Conference Proceedings ICASE 2015 FINAL

Fourth International Conference on Aerospace Science & Engineering (ICASE 2015)
ICASE - 2015
Fourth International Conference on

Aerospace Science & Engineering
September 2-4, 2015
Institute of Space Technology

Islamabad Pakistan
Proceedings
Conference Proceedings
Editors
Dr. Najam Abbas Naqvi
Mr. Raza Butt
ISBN 978-1-4673-9123-8
Printed in Pakistan
2016
Proceedings
ICASE ORGANIZING COMMITTEE

Engr. Imran Rahman
Mr. Zia Sarwar
Dr. Zafar Mohammad Khan
Dr. Muddassar Farooq
Dr. Badar Munir Ghauri
Dr. Qamar-ul-Islam
Dr. Abid Ali Khan
Dr. Ibrahim Qazi
Dr. Farrukh Chishtie
Dr. Asif Israr
Dr. Salman Ahmed
Dr. Mirza Muhammad Naseer
Engr. Ishaat Saboor
Engr. Khurram Humaiyun
Mr. Muhammad Hafeez
(Chairman)
(Secretary)
(Treasurer)
INTERNATIONAL SCIENTIFIC COMMITTEE

Dr. Leonardo Reynari , Italy
Dr Dongkai Yang, China
Dr. DDGL Dahanayaka , Sri Lanka
Dr. Ali Imran , United Kingdom
Dr. Muhammad Yusof Ismail , Malaysia
Dr. Rakhshan Roohi , Australia
Dr. Fawad Inam , United Kingdom
Dr. Tahir I. Khan , Canada
Dr. Iftikhar Ahmad , Saudi Arabia
Dr. Aquib Moin, South Africa
NATIONAL SCIENTIFIC COMMITTEE

Dr. Badar Munir Ghauri
Dr. Muddassar Farooq
Dr. Qamar-ul-Islam
Dr. Abid Ali Khan
Dr. Asif Israr
Dr. Ibrahim Qazi
Dr. Farrukh Chishtie
Dr. Syed Wilayat Hussain
Dr. Ahtezaz Qamar
Dr. Muhammad Zubair Khan
Dr. Umer Iqbal Bhatti
Dr. Aamir Habib
Dr. Khurram Khurshid
Dr. Moazam Maqsood
Dr. Fazeel Mehmood Khan
Dr. Waqas Qazi
Dr. Rizwan Mughal
Dr. Arjumand Zaidi
Dr. Abdul Haseeb
i
Proceedings
EDITORIAL COMMITTEE
Mr. Raza Butt
Mr. Waqas Ramzan
ICASE SECRETARIAT
Dr. Najam Abbas
(Secretary ICASE 2015)
Mr. Zeeshan Fareed
(Marketing and Publicity)
Mr. Raza Butt
(Media and Public Relations)
Mr. Waqas Jilani Joiya
(Logistics and Operations)
Mr. Muhammad Adeel
(Graphic Designing)
Mr. Waqas Ramzan
(Data Management)
ii
Proceedings
SPONSORS
Higher Education Commission (HEC)
National ICT R&D Fund
COMSTECH
Pakistan Atomic Energy Commission (PAEC)
Kahuta Research Laboratory (KRL)
Pakistan Science Foundation (PSF)
Vital Group
iii
Proceedings
Proceedings
PREFACE
Institute of Space Technology (IST), Islamabad, Pakistan organized the Fourth International
Conference on Aerospace Science & Engineering (ICASE) from September 2-4, 2015. This
conference is a regular biennial event to provide an international forum in which researchers,
engineers, professional and students from all over the world get a chance to interact and discuss
the latest themes and trends related with aerospace science and engineering. It provides a
platform to share experiences, foster collaborations across industry and academia, and to
evaluate emerging aerospace technologies and developments across the globe.
The success of the first three conferences in 2009, 2011 and 2013 has earned ICASE a high
standing in the domains of high performance aerospace materials, space communication
techniques, control and guidance systems, design and construction of space systems and
structures. These conferences provided an ideal opportunity for exchange of information
amongst scientists, engineers and researchers from all across the globe.
ICASE 2015 featured a diverse blend of thematic areas including Aerospace and Avionics,
Satellite Design Development and Security, Mechanical Engineering for Aerospace
Applications, Aerospace Materials Design and Engineering, Satellite Communication and Image
Processing, Global Navigation Satellite Systems, Remote Sensing & Geographic Information
Science, Astronomy and Astrophysics, Information Technology and Cyber Security, Space
Technology Awareness and Society.
A total of 110 papers were presented in the conference while 30 poster presentations were held.
There were 20 technical sessions during the conference covering the different themes and track
related with aerospace science and engineering. In addition to that, there were 15 panel
discussions, tutorials and workshops sessions in connection with conference themes. A galaxy of
30 national and International invited speakers shared their research accomplishments with the
academicians, researchers and students from all over Pakistan.
The representatives from industry and elite Research and Development organizations also
exhibited their industrial and technical paraphernalia during the conference. Extensive
deliberations and collaborations were the other significant focuses of ICASE 2015. Key
representations included National ICT R&D Fund, AIDL and the National Space Agency of
Pakistan, SUPARCO. Main sponsors of ICASE 2015 included Higher Education Commission
(HEC), PAEC, KRL, COMSTECH, Pakistan Science Foundation, National ICT R&D Fund and
the VITAL group.
Prodigious efforts were put in to publish this ICASE 2015 proceedings book. Organizing
committee, reviewers, chairs, co-chairs, data processors, proofreaders and the designers
contributed their ration remarkably in pooling the valuable research findings in a single
document. I am grateful to ICASE 2015 team for their extended efforts in making this
conference a great success. My special thanks to our sponsors for their generous financial
support in driving the research zeal amongst the researchers, scientists and the engineers
community.

Secretary
ICASE 2015
iv
Proceedings
CONTENTS
1
Junaid Anwar
A Comparison Study of Advanced state Observers for Quad rotor UAV

with Sliding Mode Control
Bushra Aijaz
Fuzzy Temperature Controller for Induction Heating
Qazi Ejaz ur Rehman
Stability and Control Solution of Quad-Copters
13
Amna Butt
Detection of Fire Hotspots dealt by Emergency First Responders in

Rawalpindi using GIS application
20
Zeeshan Khan
Prospects of Airborne Wind Energy Systems in Pakistan
25
Muhammad Amin
Integrated use of Potential Rainwater Harvesting Site for Agriculture

Using Geo-Spatial Approach
31
Muhammad Usman Saleem
38
Asad Abbas
Khazar Hayat
Urban change detection of Lahore (Pakistan) using the Thematic

Mapper Images of Landsat since 1992-2010
A Review of Fundamentals of Hyperspectral Imaging and its
Applications
A numerical study on the impact resistance of braided composites
10
Waheed Gul
Improving physical and mechanical properties of medium density fiber

board (MDF)
57
11
Abdur Rehman
64
12
Rabia Zafar
Finite Element Analysis of Tool Wear in Ultrasonically Assisted

Turning
Metamaterials in Aerospace Industry: Recent Advances and Prospects
13
Engr Numan Khan
Finite Element Simulation of Composite Body Armor
73
14
Atiq ur rehman
79
15
Malik Abid Hussain
Incorporate GNSS with Android & Improve the Search and Rescue
operation.
Geostatistical Analysis on Seismic Data over North-Western Regions of
Pakistan, Afghanistan and Eastern Regions of Iran & Tajikistan
16
Rabia Tabassum
GIS for estimating Optimized Water Demand Using Sustainable Water

Resource ManagmentFor Planned City
98
17
Ferheen Ayaz
109
18
Farzan Javed Sheikh
Optimized Threshold Calculation based on Received Signal

Characteristics for Blanking Non-linearity at OFDM Receivers
A Review on Mobile Wireless Communication Networks (0G to
upcoming generations)
19
Muhammad Altaf Khan
Intelligent Detection of Distributed Flooding Attack in Wireless Mesh

Network
120
20
Aamir Nawaz
Touch panel Based Restaurent Automation Using Zigbee Technology
126
21
Ferheen Ayaz
Introducing space plant biology to students through hands-on activities

using clinostat
131
22
Faizan Muhammad
The Political and Economic Feasibility of Current Space Resource

Management Policies
136
23
M Sohail Shahid
147
24
M. Saad Sohail
Feasibility Study to Install Fire Fighting Equipment on a Cargo

Helicopter
Design and Optimization of S-band Wilkinson Power Divider for
Transceiver Applications
25
Mateen Tariq
157
26
Taimoor Zahid
27
Najam ud Din Ahmad
28
Syed Jahanzeb Hussain Pirzada
Optimize Manufacturing of unidirectional carbon prepregs for space

Applications
Electrical Power Conditing Unit Design for Space Qualified C Band
Receiver Geo Satellite Applications
DSP based Electro Hydraulic actuator control with irreteraceable
feedback error.
Design for Test Approach using FPGA for BPSK Modem
29
Taimoor Zahid
Design of a Fuzzy Logic Water Level Controller
175
30
Gohar Ali
The use of Nuclear reactor in Space Applications:Propulsion and Power

Concept
181
44
50
69
84
114
154
160
165
171
Proceedings
31
Muhammad Aamir
Impact of Thermal Aging on Microstructure and Mechanical Properties

of high Sn Content, Sn-Pb Solders
187
32
Madni Shifa Ullah Khan
Effect of aryl diazonium salt functionlization on the electrical properties

of MWCNTs and MWCNTs/CF reinforced polymer composite
192
33
Sania Nazir
Design of C band Slotted Waveguide Array antenna with high

Impedance Bandwidth and improved Reflection Coefficient
200
34
Muhammad Nauman Hussain
Risk Areas Mapping and Identification of Hotspots on the RoadNetwork of Lahore
203
35
Muhammad Arslan
Analysis of Recent Drought Based on NDVI and Meteorological Data
208
36
Ali Jan Hassan
Assessment of Urban Growth of Karachi: From A Tiny Town to A Meta

City of the World
211
37
Abdur Raqeeb Gaziani
JUPITER, The Gas GIANT
218
38
Shakeel Ahmad Waqas
219
39
M Ameer Umar Malik
40
M Shahan Qamar
Identification of Post Disaster Scenario Using Double Threshold Energy

Detection
Design and Analysis of Magnetic MEMS Accelerometer for Inertial
Navigation
Feasibility Assessment of Running JP-8 Fuel in Diesel Engine
41
Artificial intelligence reboot
236
42
Izhar
243
43
Sundus Najib
An Intelligent Approach for Edge Detection in Noisy Images Using

Fuzzy Logic
A Survey of Active ITU-R P-Series Propagation Models
44
Anwar Ul Haque
An Experimental Study To Evaluate the Effect of Strut and Fairing
255
45
M Tanveer Iqbal
Fairing Separation Dynamic Analysis Using Analytical Approach
261
46
Saqib Alam
Launch Vehicle Control based on Dynamic Inversion with Sliding

Mode Neural Network augmentation.
265
47
Syed Wasif Ali Shah
Design and Development of Low Cost Motor Drive for Hub Wheel
based Electric Vehicles.
271
48
Shahid Karim
Preparation, Structure and Dielectric/Piezoelectric Properties of BiScO 3PbTiO3-Pb(Mn1/3Nb2/3)O3 high temperature piezoelectric ceramics
275
49
Abdur Rasheed
A Comprehensive Study On QoS For Mobile Ad Hoc Network
282
50
Muhammad Shoaib
Comparison of Maximum Likelihood Classification Before and After

Applying Weierstrass Transform
289
51
Hira Fatima
Spatio -Temporal Analysis of Shoreline Changes along Makran Coast

Using Remote Sensing and Geographical Information System
296
52
Sadaf Javed
316
53
Iqra Basit
Influence Analysis of Minerals on Drinking Water Quality Around

River Jhelum
Selection of the optimal interpolation method for groundwater quality
54
55
M Tasawer Hussain
Zehra Ali
Thermal Design and Analysis of PNSS-1 Satellite

Optimized Threshold Calculation based on Received Signal
Characteristics for Blanking Non-linearity at OFDM Receivers
334
340
56
Naveed Riaz
345
57
IEEE Publications
Measurement & Testing Techniques of Performance Parameters for

Electric Servo Actuators
LIST of 27 IEEE submitted Papers
vi
222
229
249
325
348
Proceedings
Nonlinear Observers for Closed loop Sliding Mode

Control of Quadrotor UAV
Junaid Anwar, Fahad Mumtaz Malik, Muhammad Bilal Khan
College of Electrical and Mechanical Engineering, National University of Sciences and Technology, Pakistan
and asymptotic convergence. Researchers have contributed in

the development of their idea towards high gain observer [17]
[20].High gain observers performance degrades in the presence
of external noise and is shown in simulation section of this
paper, due to which we have to look for an observer which is
robust to sensor noise.
The sliding mode observers are widely used because of their
prominent features like finite time convergence, robustness to
sensor noise and un-certain estimation [21], [22].Asymptotic
second order sliding mode observers were also developed but
they require proof of separation principle.High accuracy and
reduction in chattering are the main features of second order
sliding mode compared to the classical first order motion.
Recently a class of second order sliding mode observer is
introduced so called super-twisting observer [23] for second
order mechanical system which include quadrotor too. Supertwisting observers can reconstruct the states if the perturbation
is of relative degree two, or reconstructs the perturbation itself,
when it is of relative degree one in finite time.
Aim of this paper is to compare the performance of both observers namely high-gain observer and super-twisting sliding
mode observer under same set of perturbations, uncertainties
and noise, so that each observer can exhibit its characteristics
for the quadrotor system. Real time estimation always require
knowledge of the pros and cons of the observer relative to the
particular system, so that best observer can be deployed for
real time estimation of states. So there is a need to explore
the type of observers which are application specific.
In the following section model of the quadrotor is developed
and presented. In section-III controller design is presented. In
section-IV observers are designed for the quadrotor model. In
section-V numerical simulation is given and finally in sectionV1 the conclusions are given.
AbstractThis paper deals with the performance comparison

of Sliding mode observer with super-twisting algorithm (STSMO)
and High gain Observer (HGO) for a remotely controlled quad
rotor UAV. Under the restriction that inertial co-ordinates and
attitude angles are available for measurement while angular
and linear velocities are estimated. This paper is solved in
two steps for each observer. First the observer (HGO and
STSMO) is designed and then in second stage a second order
(2-SMC) technique is being applied on the basis of estimated
states to design controller(for which systems is portioned into
fully and under-actuated sub-systems).Simulations results shows
the performance comparison of both observers under the same
control scheme.
I. I NTRODUCTION
More recently, a growing interest in the UAV has been
shown by industry and academia [1][7].The vital and potential use of flying robots for civil as well as military applications
are attracting the industries and the academia community. The
feature of flying in narrow space and vertical takeoff and landing (VTOL) made quadrotor unique relative to other mobile
robots and conventional aircrafts. The quadrotor is an under
actuated system with six outputs and four inputs, they are
owed to carry out the tasks ranging from surveillance to rescue
mission but the challenges behind the control of quadrotor
aerial vehicle like un-stability and highly nonlinear behavior
are the major source of attraction and many control approaches
to deal with quadrotor dynamics have been presented so far
[8][14][14-20].
This paper deals with the development of 2-sliding mode
control scheme that can cater for the model uncertainties,
external disturbances and the chattering phenomenon. Nonavailability of states is a major constraint towards accomplishment of any control scheme, using sensor for each state is
also not feasible due to space limitations and high cost of the
sensors. Even with the availability of all states system model
generally shows parametric mismatch with respect to the
real time environment. These model imperfections, un-certain
initialization and sensor errors also degrades the performance
of the controller. The solution for that is to use state observers,
to estimate the states in real time, Luenberger [15] proves to be
good in the state estimation but these model based observers
fail when the system parameters keep on changing with the
time. Least square and recursive least square (RLS) are also
not able to work on highly nonlinear system such as quadrotor.
A high gain observer was first introduced by Khalil and
Esfandiaro [16] for the design of output feedback controllers
II. DYNAMIC MODELING

A quad rotor UAV is a highly nonlinear dynamical system
and its modeling it is not an easy task due to under actuation. It consists of two pairs of rotors which are moving in
opposite directions to provide the collective thrust as shown
in following fig 1. There are four inputs to this system.
The input U1 is the sum of thrusts provided by individual
rotors. The pitch movement is obtained by changing speeds
of rotors 2 and 4. Similarly the roll movement is achieved by
varying speeds of rotor 1 and 3. These two former operations
should be performed while keeping the total thrust constant
otherwise system may lose altitude and crash. The roll and
pitch movements are controlled by using inputs U2 and U3
respectively. The yaw movement occurs due to difference
between torques of the two pairs of rotors. This movement
is stabilized by using input U4 .
Proceedings
The crude and approximate model of quad rotor UAV from

above set of equations can be written as follows
= v
v = ge3 + Re3
b X 2
i
ms
R = Rw
J w = w Iw
Fig. 1.
cos cos
cos sin
sin
Quad Rotor UAV Free Body Diagram
cos sin + sin sin cos

cos cos + sin sin sin
cos sin

Jr w e3 i +
whereTd is some modeling co-efficient, e3 is a vector

0 0 1 ,r is the rotor co-efficient, is the position vector,
R is the transformation matrix, w
is the skew-symmetric
matrix, is the rotor speed, I is the inertial tensor matrix,
Jr is the rotor inertia, while Jp and Jm are the propeller and
motor inertia respectively and b is the thrust co-efficient.
The torques applied to the quad rotors axis is the difference
between the torques provided by the rotors on the other axis.

!
lb 24 22
lb 23 21
=

2
d 2 + 24 23 21
The rotor inertia consists of motor inertia, propeller inertia

and negligible reversing gearbox inertia and is given by the
following equation
Let us consider an inertial frame of reference E and a body

fixed frame B as shown in above figure. The transformation
between E and B is provided by a matrix R which is given by
the following equation.
xn !
yn
=
zn
sin sin + cos cos sin !

sin sin cos cos sin
cos cos
Jr = Jp Jm r
Now the complete six degrees of freedom model is given
by the following system of equations:-
xb !
yb
zb
U1
m
U1
y = cos sin sin sin cos
m
cos cos
z = g +
U1
ms
Iy Iz
Jr
l
=
r + U2
Ix
Ix
Ix
I
I
J
l
z
x
r
=
r + U3
Iy
Iy
Iy
Ix Iy
C
=
+ U4
Iz
Iz
x
= cos sin cos + sin sin
xn !
yn
=R
zn
xb !
yb
zb
The Newton-Euler formalism is used to present the dynamics of quad rotor UAV. The Newtons laws of motion when
applied to a rigid body in the presence of external forces and
torques are given by following set of equations
ms I33
O
O
I
V
w
!
+
w ms V
w Iw
!
=
where ms is the mass of the quadrotor, V vector of xyz,

w
be the vector
T of , and ,I represents a inertia vector
Ix Iy Iz
across x, y and z respectively. is torque
vector include roll
T torque,pitch torque and yaw torque and
F = 0 0 U1 .To convert the above equations in inertial
frame we use transformation matrix to get the following
equations
(1)
where C is the proportional constant.The first term on right

hand side of first dynamical equation is the gyroscopic effect
caused by the rotation of the rigid body and the second term
is due to the propulsion effect. The system inputs are U1 , U2 ,
U3 and U4 .
The inputs are given by the following equations
U1 = b 22 + 24 23 21

U2 = b 24 22

U3 = b 23 21
= v
F
ms
R = Rw
v = R
J w = w Jw +
U4 = d 22 + 24 23 21
= d 4 + 2 3 1
To make the model more realistic especially in forward

flight we include the hub forces, rolling moments and variable
aerodynamics coefficients [24].
The hub force is the resultant of horizontal forces acting on
all blade elements.
Proceedings
is given by the following set of equations:x

= cos sin cos + sin sin
4
1 X
x
Hxi K1
m i=1
ms
y = cos sin sin sin sin
H = CH A Rrad
U1
m
2
U1
m
1 X
y
Hyi K2
m i=1
ms

cos cos
z
U1 + Fgr z K3
ms
ms
4
X
Iy Iz
Jr
l
h
=
r + U2
Hyi +
Ix
Ix
Ix
Ix i=1
z = g +
where CH is the hub coefficient, is the air density and A is

the propeller disk area and Rrad is the propeller radius and
is the speed of the respective propeller.Additionally the drag
moment Q is the moment about the rotor shaft caused by the
aerodynamic forces acting on the blades. In fact drag moments
determine the power required to spin the rotors. It is given by
the following equation
4
(1)i+1 X
Rms xi lK4
Ix
Ix
i=1
(2)
4
Iz Ix
Jr
l
h X
=
r + U3
Hyi +
Iy
Iy
Iy
Iy i=1
4
(1)i+1 X
Rms yi lK5
Iy
Iy
i=1
2
Q = CQ A Rrad Rrad
4
Ix Iy
C
h X
=
+ U4 +
Hyi +
Iz
Iz
Iz i=1
where CQ is the drag coefficient.The rolling moment of

a propeller exists in forward flight when advancing blade
is producing more lift than the retreating blade. It is the
integration over the entire rotor of the lift of each section
acting at a given radius and is given by following equation.
4
X
l
Hx2 Hx4 + Hy3 Hy 1
Qi yi
Iz
i=1
lK6
Iz
III. CONTROLLER DESIGN

, ,
, ,
T and U =
Let X = x, x,
y, y,
z, z,
, ,

T
U1 , U2 , U3 , U4
be the state and control input vectors respectively. The equation set (1) with the addition of friction
term and ground effect term for altitude can be written in state
space representation such as:
2
Rm = CRm A Rrad Rrad
Where CRm is the rolling moment coefficient.Furthermore the

UAVs operating near the ground (approximately at half rotor
diameter) experience thrust augmentation due to better rotor
efficiency. This is related to a reduction of induced airflow
velocity. This is called Ground Effect. The following equation
represents the ground effect near the surface. It is assumed
that ground effect acts on the UAV when the UAV is below a
certain altitude,zo .
x 1 = x2
U1
x 2 = cos x7 sin x9 cos x11 + sin x7 sin x11
m
x2
K1
ms
(3)
x 3 = x4
Fgr z =
A
A
z + zcg )2
zo + zcg )2
U1
x 4 = cos x7 sin x9 sin x11 sin x7 sin x11
m
x4
K2
ms
0 < z zo
(4)
x 5 = x6
After incorporating the above effects and the friction terms,

we obtain a more realistic model of the quad rotor UAV which
x 6 = g +

cos x7 cos x9
x6
U1 + Fgr z K3
ms
ms
(5)
n
Iy Iz
Jr
= S1 d

r
Ix
Ix
o
l
U2 + K4 l + 1 d
Ix
Ix
x 7 = x8
x 8 = x10 x12
Iy Iz
Jr
l
x8
r x10 + U2 lK4
Ix
Ix
Ix
Ix
(6)
Iz Ix
Jr
l
x10
r x8 + U3 lK5
Iy
Iy
Iy
Iy
(7)
To make it negative definite choice of input is as:

Ix n
Iy Iz
Jr
U2 =
d

r
l
Ix
Ix
o

+ K4 l + 1 d + k1 sat S1 + k2 S1
Ix
where
k1 , k2 > 0
x 9 = x10
x 10 = x10 x12
x 11 = x12
x 12 = x10 x8
(8)
Ix Iy
C
x12
+ U4 + lK6
Iz
Iz
Iz
Similarly in the same way surface for subsystem (5) comes out
to be the linear combination of position and velocity tracking
errors of z state.

S2 = zd z + 2 zd z
where 2 > 0
The term represents the ground effect near the surface. It is

assumed that the ground effect acts on the UAV when the UAV
is below a certain altitude. The goal is to design a second
order sliding mode control which is done in two steps. 1)
Choice of sliding surface w.r.t tracking error e. 2) Design
of Lyapunov function that guarantees negative definiteness
so that asymptotic convergence is achieved. Closed loop
control system dynamics become insensitive to modeling error,
perturbation signals and parameter variation as a by-product of
sliding mode control (SMC). Control efforts are calculated by
the help of Lyapunov analysis and hence guarantee asymptotic
convergence. As quadcopter has 4 inputs while number of
variables to be controlled are more than four hence overall it
is an under actuated system but we can portioned that system
into two parts namely fully actuated part and under-actuated
part and then designing the control for each part of the system
independently. Therefore control is also portioned into two
sub-system.
And in the same way by the Lyapunov analysis of surface the

control comes out to be
n

ms
z
U1 =
zd + g + K3
+ 2 zd z
cos cos
m
o s

+ k2 sat S2 + k4 S2 + Fgr
where
B. Control for Under-actuated subsystem

Under-actuated subsystem composed up of x
, y, and
subsystems. The Choice of sliding surface for the subsystem (3) and (7) comes out from the Lyapunov analysis as:
e2x
e2
+ where ex = xd x
2
2
= ex ex + e e

= ex xd x + e d
V =
Fully-actuated subsystem composed up of z and subsystems (5) and (6) respectively. Choice of sliding surface for the
subsystem (6) comes out from the Lyapunov analysis as:
V
V
V
V
V
So to make V negative definite

x = xd + 3 xd x and
e = d
V < 3 xd x
So to make V negative definite

= d + 1 d
2
4 d
and
e = d
= d + 4 d
2
where
3 , 4 > 0
Hence Surface S3 will be

S3 = xd x + xd x + d + d
Hence Surface S1 will be
Let the Lyapunov function be
S1 = d + 1 d
S32
2
V = S3 S3
n

Iz Ix Jr
= S3 xd x
+ 3 xd x + d
+ r
Iy
Iy
o

l
l
U3 + K5 + 4 d
Iy
Iy
V =
Let the Lyapunov function be

S12
2
V = S1 S1
n
o
= S1 d 1 d
k3 , k4 > 0
Where
A. Control for fully actuated subsystem
e2
=
where
2
= e e

= e d
Proceedings
V =
where
1 > 0
To make it negative definite choice of input is as:

Iy n
Iz Ix
Jr
U3 =
d
r lK5 + 3 x d x
l
Iy
Iy
Iy

+ 4 d + xd x
+ k5 sat(S3 ) + k6 S3
where
IV. OBSERVER DESIGN

Two types of nonlinear observers are implemented for the
quad rotor system with the same control scheme i.e. 2-Sliding
mode Control. Observability is ensured by [25] for each block
of equation from (3)(8) separately.
k5 , k6 > 0
Similarly in the same way surface for subsystem (4)

and eqref3f comes out to be the linear combination of position
and velocity tracking errors of two states i.e. y and .

S4 = yd y + 5 yd y + d + 6 d
where
A. High Gain Observer

HGO is basically an approximate differentiator. This observer works well for a wide class of nonlinear systems and
leads to recovery of the performance achieved under state
feedback. Implementation of this observer is quite simple
because it needs less computational effort with an additive
advantage of this observer is that its performance doesnt
degrade with the presence of model uncertainties in the plant.
High gain observer is an asymptotic observer and dynamics of
this observer can be made arbitrarily fast through epsilon and
gains alphas. Separation principle theorem doesnt need to
be proved and high gain observer can be designed separately
from the controller. The HGO is applied to multiple input and
multiple output system as:
5 , 6 > 0
and in the same way by the Lyapunov analysis of surface S4

the control U4 comes out to be

Ix Iy
Iz
U4 =
d
lK6 + 5 yd y
c
Iz
Iz

6 d + yd y + k7 sat(S4 ) + k8 S4
where
Proceedings
k7 , k8 > 0
U1 is the control input of z, U2 is the control input of roll,U3

is the control input of pitch and U4 is the control input
of yaw while motion in x and y direction is produced by
help of control inputs of roll, pitch and z. By the help of
U1 , U2 , U3 and U4 the desired trajectories are achieved and
tracking errors are reduced to zero asymptotically, by virtue
of Sliding mode controller.by keeping the roll and pitch angles
to zero controller robustly stabilize the UAV and move it to the
desired position with a desired yaw angle. The control scheme
is developed and implemented independent of observer and is
shown in the block diagram The controller is designed by
keeping in view the mathematical model of the quad rotor as
given in Section-II without any effects except friction and the
ground effect but that U1 , U2 , U3 and U4 are capable enough to
tackle not only Rolling Moments, Drag moments, Gyroscopic
effects, Hub forces but also retain its performance which is
being shown in the simulations section in this paper.

x = Ax + B x, u
The HGO, then, is given by

= A
x
x + B0 x
, u + H y C x
y = Cx

W here H = blockdiag H1 , H2 , H3 , H4 ,
"
Hi =
11
22
#
i = 1, 2, 3, 4
= positive constant, and constant parameter ji are

obtained from a Hurwitz polynomial, j=1,2
s2 + 11 s + 22 = 0
The HGO for the quadcopter system is designed in blocks
as [26] i.e six HGOs are designed for each block of equation
from (3)(8) separately. For equation (3) HGO is implemented
as
"
#
"
#
"
#
0 1
0
1
With A =
,B=
and C =
and
0 0
1
0
0 = cos x7 sin x9 cos x11 + sin x7 sin x11
U1
x2
K1
m
ms
and H1 is designed as in the aforementioned equation. Similarly for equation (4), (5),.....(8).HGOs are given in the same
way as for equation (3).constant gains are enlisted in the table
1 in simulation section of this paper.
Fig. 2. Controller designed independent of the observer
state feedback after 12 seconds while x state a bit earlier as

compared to other states, both observers are initialized with
same initial conditions, so that performance can be compared
in a proper way and all the observers parameters are listed
in the table 1 the performance of HGO is slightly better than
STSMO in tracking as HGO estimates desired values a bit
earlier as compared to STSMO. The chattering problem is
intelligently avoided in the sliding mode control by using
continuous approximation to the sign function. This makes this
approach applicable in real applications. As the control laws
are developed for set of equations (1) but implemented on
set of equations (2) which include different kind of effects as
mentioned in section-II, similarly the model used for observer
is based on set of equations (1) and observer giving estimate
for set of equations (2) which is quadrotor model with ground
effects, drag moment, rolling moment and pitch moment.
B. Sliding mode Observer with super-twisting Algorithm

One of the best sliding mode observer which offers a
finite time reaching time [27], [23] and which can be used
for sliding mode based observation is the super-twisting observer.Separation principle theorem is trivial in this case too
and the Super-twisting sliding mode observer can be designed
separately from the controller. The finite time convergence
property of sliding mode observers is usually suitable in the
scheme of observation and for the purpose of observer-based
controller design for nonlinear systems Super-twisting sliding
mode observer has the form

1 = x
(9)
x
2 + |x1 x
1 |1/2 sat x1 x
1

2 = f x1 , x
x
2 , u + sat x1 x
1
Proceedings
(10)
Taking x
1 = x1 x
1 and x
2 = x2 x
2 we obtain the error
equations as

x
1 = x
2 |
x1 |1/2 sat x
1

x
2 = F t, x1 , x2 , x
2 sat x
1
where,

F x1 , x2 , x
2 = f x1 , x2 , u f x1 , x
2 , u + x1 , x2 , y
is used for perturbations.For the bounded states, existence
of a constant is ensured such that

|F x1 , x2 , x
2 |< f +
Observer designed by equation (9) and (10) takes into account of partial knowledge of system dynamics while setting
parameters and and hence more accurate. The full order
Super-twisting Sliding mode observer for equation (3) is given
as

1 = 1 + x
x
1 |x1 x
1 |1/2 sat x1 x
1

1
x
2
x 2 =
cos sin cos + sin sin U1 K1
ms
ms

1 sat x1 x1
Fig. 3. HGO and STSMO tracking of desired values
1 and 1 are designed by the help of aforementioned inequality as [27]. Similarly for equations (4), (5),....., (8) Supertwisting sliding mode observers are implemented in the same
way as for equation (3), while gains are given in the table.1
in simulation section of this paper.
V. SIMULATION STUDY
A. Closed loop Simulation with model uncertainties and without noise for HGO and STSMO
Fig. 4. HGO and STSMO tracking of desired values
Simulation results for observer-based controller of the

quadrotor are shown in the fig.3 and fig.4 for both observers
HGO and STSMO. Now under output feedback, controller
is in conjunction with HGO and STSMO separately and is
using all the states of the observer. The results in the fig.3
shows that both High gain observer and as well as Supertwisting sliding mode observer recovers the performance of
B. Closed loop Simulation with noise and model Uncertainties

for HGO and STSMO
A constant noise of 0.1 value is added in the output of
the system in each of six states and the results obtained after
simulation for each observer are shown in the fig.5 and fig.6
Proceedings
which indicated that in the case of STSMO no effect on

the observers estimated states while on the other hand HGO
estimates deviated by the same amount as disturbance added
which opens an era of coupling integral control scheme with
HGO to eliminate the steady state error
Fig. 8. Control Effort with HGO
Fig. 5. HGO Estimated values under constant sensor noises
Fig. 9. Control Effort under the state feedback

TABLE I
THE NOMINAL PARAMETERS AND THE INITIAL CONDITIONS
OF THE OBSERVER AND THE SYSTEM FOR THE QUADROTOR
MODEL
Fig. 6. STSMO Estimated values under constant sensor noise
C. Effect of observer scheme on Control effort required

Control effort is greatly affected by the effects namely
drag moment, rolling moments, pitch moments and hub forces
which are included in the system model but not included
in any of the observer, and by the type of observer used.
Simulations results in fig.7 and fig.8 shows that HGO required
larger control effort in transient phase as compared to STSMO
and fig.9 shows that over all with the conjunction of observer
in the closed loop model control effort in transient phase in
considerably increased
Variable
Value
ms
l
Ix = Iy
Iz
lr
K1 , K2 , K3
K4 , K5 , K6
g
b
C
k1 , k3
k2 , k4
k5 , k7
k6 , k8
1 , 3 , 5
2 , 4 , 6
zcg
1.1
0.21
1.22
2.2
0.2
0.1
0.12
9.81
5
1
0.8
2
0.5
5
2
6
0.1
4
2,1,4,4
6,9,10,25
0.9
1
5
2,1,6,9
1 , 2 , 3 , 4
5 , 6 , 7 , 8
9 , 10 , 11 , 12
Units
Initial Condition
High Gain
Observer
Super Twisting Sliding

Mode Observer
kg
m
x
1 (0)
x
2 (0)
x
3 (0)
x
4 (0)
x
5 (0)
x
6 (0)
x
7 (0)
x
8 (0)
x
9 (0)
x
10 (0)
x
11 (0)
x
12 (0)
1
2
0.6
-2
2
1
-1
1
0.5
1
1.3
3
1
2
0.6
-2
2
1
-1
1
0.5
1
1.3
3
N s2 /rad
N s2 /rad
N s2 /rad
N s/rad
N s/rad
m/s2
N s2
VI. CONCLUSION
This paper has presented a comparison study of nonlinear
observers, including high gain observer and super-twisting
sliding mode observer in conjunction with the 2-Sliding mode
controller for the quadrotor system under external disturbances
and model uncertainties. The highgain observer can cater for
the model uncertainties but not the external disturbance while
the super-twisting sliding mode observer not only cater for
the model un-certainities but can also performs well under
external disturbances (sensor noise). The second important
result regarding initialization of high-gain observer is that it
Fig. 7. Control Effort with STSMO
*Initial conditions for the quadrotor system are set to zero

deliberately for evaluating performance comparison of both
observers.
doesnt allow random initialization unless gains are adjusted on

the other hand super-twisting sliding mode observer provides
flexible environment in initialization.
None of these observes is computationally onerous, but supertwisting sliding mode observer utilizes the knowledge of
system partially [27] as compared to high-gain observer which
just rely on the Hurwitz polynomial and the tuning parameter
[26] so this fact also indicates the practical applicability of
super-twisting sliding mode observer in the cases where model
uncertainties are bounded as it gives finite time convergence as
compared to asymptotic convergence in the case of high gain
observer which are favourable in the environment where model
uncertainties are present or parameters are time varying in
those conditions these filters are preferable than super-twisting
sliding mode observer only. This effort will be a good starting
point to explore super-twisting sliding mode observers and to
compare them with other observers of its breed (higher order
sliding mode observers).
Proceedings
[14] A. Benallegue, A. Mokhtari, and L. Fridman, Feedback linearization

and high order sliding mode observer for a quadrotor uav, in Variable
Structure Systems, 2006. VSS06. International Workshop on. IEEE,
2006, pp. 365372.
[15] D. G. Luenberger, Observers for multivariable systems, Automatic
Control, IEEE Transactions on, vol. 11, no. 2, pp. 190197, 1966.
[16] F. Esfandiari and H. K. Khalil, Output feedback stabilization of fully
linearizable systems, International Journal of control, vol. 56, no. 5,
pp. 10071037, 1992.
[17] A. N. Atassi and H. K. Khalil, A separation principle for the stabilization of a class of nonlinear systems, IEEE Transactions on Automatic
Control, vol. 44, no. 9, pp. 16721687, 1999.
[18] A. Isidori, A remark on the problem of semiglobal nonlinear output
regulation, IEEE transactions on Automatic Control, vol. 42, no. 12,
pp. 17341738, 1997.
[19] Z. Lin and A. Saberi, Robust semiglobal stabilization of minimumphase input-output linearizable systems via partial state and output
feedback, Automatic Control, IEEE Transactions on, vol. 40, no. 6,
pp. 10291041, 1995.
[20] W. J. Rugh, Linear system theory. prentice hall Upper Saddle River,
NJ, 1996, vol. 2.
[21] J. Barbot, M. Djemai, and T. Boukhobza, Sliding mode observers,
Sliding Mode Control in Engineering, vol. 11, 2002.
[22] C. Edwards, S. K. Spurgeon, and C. P. Tan, On the development and
application of sliding mode observers, in Variable Structure Systems:
Towards the 21st Century. Springer, 2002, pp. 253282.
[23] J. Davila, L. Fridman, and A. Levant, Second-order sliding-mode
observer for mechanical systems, IEEE transactions on automatic
control, vol. 50, no. 11, pp. 17851789, 2005.
[24] M. Becker, R. C. B. Sampaio, S. Bouabdallah, V. d. Perrot, and
R. Siegwart, In-flight collision avoidance controller based only on
os4 embedded sensors, Journal of the Brazilian Society of Mechanical
Sciences and Engineering, vol. 34, no. 3, pp. 294307, 2012.
[25] I. Khan, A. Bhatti, Q. Khan, and Q. Ahmad, Sliding mode control
of lateral dynamics of an auv, in Applied Sciences and Technology
(IBCAST), 2012 9th International Bhurban Conference on. IEEE, 2012,
pp. 2731.
[26] H. K. Khalil and J. Grizzle, Nonlinear systems. Prentice hall New
Jersey, 1996, vol. 3.
[27] Y. Shtessel, C. Edwards, L. Fridman, and A. Levant, Sliding mode
control and observation. Springer, 2014.
R EFERENCES
[1] M. Chen and M. Huzmezan, A simulation model and h (loop shaping
control of a quad rotor unmanned air vehicle. in Modelling, Simulation,
and Optimization, 2003, pp. 320325.
[2] S. Bouabdallah, P. Murrieri, and R. Siegwart, Design and control
of an indoor micro quadrotor, in Robotics and Automation, 2004.
Proceedings. ICRA04. 2004 IEEE International Conference on, vol. 5.
IEEE, 2004, pp. 43934398.
[3] B. Herisse, T. Hamel, R. Mahony, and F.-X. Russotto, Landing a vtol
unmanned aerial vehicle on a moving platform using optical flow,
Robotics, IEEE Transactions on, vol. 28, no. 1, pp. 7789, 2012.
[4] L. Derafa, A. Benallegue, and L. Fridman, Super twisting control
algorithm for the attitude tracking of a four rotors uav, Journal of the
Franklin Institute, vol. 349, no. 2, pp. 685699, 2012.
[5] A. Tayebi and S. McGilvray, Attitude stabilization of a vtol quadrotor
aircraft, Control Systems Technology, IEEE Transactions on, vol. 14,
no. 3, pp. 562571, 2006.
[6] L. Besnard, Y. B. Shtessel, and B. Landrum, Quadrotor vehicle control
via sliding mode controller driven by sliding mode disturbance observer,
Journal of the Franklin Institute, vol. 349, no. 2, pp. 658684, 2012.
[7] M. Bouchoucha, S. Seghour, and M. Tadjine, Classical and second
order sliding mode control solution to an attitude stabilization of a four
rotors helicopter: From theory to experiment, in Mechatronics (ICM),
2011 IEEE International Conference on. IEEE, 2011, pp. 162169.
[8] E. Altug, J. P. Ostrowski, and R. Mahony, Control of a quadrotor
helicopter using visual feedback, in Robotics and Automation, 2002.
Proceedings. ICRA02. IEEE International Conference on, vol. 1. IEEE,
2002, pp. 7277.
[9] E. Altug, J. P. Ostrowski, and C. J. Taylor, Quadrotor control using
dual camera visual feedback, in Robotics and Automation, 2003.
Proceedings. ICRA03. IEEE International Conference on, vol. 3. IEEE,
2003, pp. 42944299.
[10] T. Madani and A. Benallegue, Control of a quadrotor mini-helicopter
via full state backstepping technique, in Decision and Control, 2006
45th IEEE Conference on. IEEE, 2006, pp. 15151520.
[11] , Backstepping sliding mode control applied to a miniature quadrotor flying robot, in IEEE Industrial Electronics, IECON 2006-32nd
Annual Conference on. IEEE, 2006, pp. 700705.
[12] P. Castillo, P. Albertos, P. Garcia, and R. Lozano, Simple real-time
attitude stabilization of a quad-rotor aircraft with bounded signals, in
Decision and Control, 2006 45th IEEE Conference on. IEEE, 2006,
pp. 15331538.
[13] N. Metni, T. Hamel, and F. Derkx, Visual tracking control of aerial
robotic systems with adaptive depth estimation, in Decision and Control, 2005 and 2005 European Control Conference. CDC-ECC05. 44th
IEEE Conference on. IEEE, 2005, pp. 60786084.
Proceedings
Fuzzy Temperature Controller for Induction

Heating
Bushra Aijaz
Rahema Kaleem
Naeema Saeed
Dept. of Electrical Engineering

Bahria University
Karachi, Pakistan
bushra.aijaz@bimcs.edu.pk
Dept. of Electronic Engineering

NED University of Engg & Tech
Karachi, Pakistan
rahemakaleem@gmail.com
Dept. of Electronic Engineering

NED University of Engg & Tech
Karachi, Pakistan
naeemasaeed991@hotmail.com
The system consists of DSK DSPC6713 for generating

SPWM pulses, isolation circuit inverter, transformer (step-up)
and temperature sensing circuit. SPWM pulses are fed to gate
drivers and then to inverter. The output is then fed to step up
transformer to obtain the desired output level which is then
fed to the load for heating. Controlling of temperature is done
by controlling the operating frequency of inverter.
Abstract this paper focuses on Fuzzy Temperature Controller

(FTC) for induction heating, which aims to provide a precise and
intelligent temperature controller. Induction heating is a non
contact process and so it is safer. The heating all depends on the
larger number of Cu and Eddy losses and thus faster. Variable
frequency Inverter is used to achieve temperature control for
induction heating coils. By controlling the output frequency of
inverter, the temperature of the load is controlled. Fuzzy logic is
implemented for overall controlling. FTC is modelled on DSP
Starter Kit DSKC6713. All programming is written in Clanguage. The model of this idea has been designed and tested
using LabVIEW 8.5.
Keywords Fuzzy Logic, Induction Heating, DSP, Variable
Frequency Inverter.
The rest of the paper is organized as follows: in Section II, we

describe the inverter in brief. Fuzzy Logic Controller is
described in Section III, SPWM is generated in Section IV,
Temperature Sensors are interfaced in section V. Induction
Heating is discussed in Section VI and the paper is concluded
in Section VII.
I. INTRODUCTION
Precise temperature controlling and fast heating processes
are integral part of industries. The industries require a new
modern technique which can handle temperature controlling
with more accuracy and precision as the controllers currently
used are slow and inaccurate plus they are unsuitable for nonlinear measurements.
The block diagram of the system is shown in Figure 1
Project Block Diagram.
II. INVERTER
This section focuses on DC to AC inverter. The purpose is to
efficiently convert a DC power source to a high voltage AC
source, which will be required to drive the load. This is
achieved by first converting low voltage DC power source to
high voltage DC power source and then the HV DC power to
AC power source using sinusoidal pulse width modulation
technique, the output of which is 220Vac.
As the induction coils operate at much higher frequencies so a
high frequency inverter is needed. To accomplish this, the
converter is designed using a full-wave rectifier. Smoothening
capacitors are connected at the output of the full-wave
rectifier to convert rectified pulsated output to smooth DC
output voltage. This output is then given to H-bridge that
gives AC square wave of 220V. The circuit and output is
shown in Figure 2.
Figure 1 Project Block Diagram
Proceedings
Figure 3Membership functions for input Fuzzy variable T
Figure 2 Design and Simulation of H- Bridge
To minimize the power loss and to ensure high switching

speeds, N-channel MOSFETs IRF840 are used in H- Bridge.
MOSFET gate drivers are used and controlled through SPWM
signals coming from the DSPC6713. The values for
bootstrapping capacitors and diodes are calculated using
Equation 1.
Figure 4 Membership functions for input Fuzzy variable T/t
The output of H-Bridge gives 220 Vac whose frequency can be

varied to more than 200Hz.
III. FUZZY LOGIC CONTROLLER
Fuzzy logic is doing all sort of controlling. The logic was
first proposed by Zadeh in 1965.
The Fuzzy logic is a linguistic logic. It is similar to crisp
logic but with the difference is that crisp logic has only two
levels of decision either 0 or 1 but fuzzy logic can have levels
in between them, which makes the logic precise and close to
ideal behaviour.
Figure 5 Membership functions for output Fuzzy variable F
B. Fuzzy Inference
A set of rules is defined for controlling purpose. In this
project, Fuzzy logic consists of 5 levels of decision for both
the inputs and output. So the rule set comes up with 25 levels.
The Table 1demonstrates the set of rules for F.
T
T/t
Fuzzy
Controller
Table 1 Fuzzy Rule Base
Delta T
Delta (T/t)
NB
NS
Z
PS
PB
The fuzzy structure is based on following three steps;

A. Pre-processing
As shown above; Error and rate-of-error are the inputs to
the fuzzy controller, whereas error is the difference of set
temperature and Feedback temperature. The inputs are
properly assigned with their membership () functions (after
observing temperature rise and fall graph described in section
V). Figure 3 and Figure 4 shows the function assignments of
the two inputs T and T/t and Figure 5 shows the
functions of output F.
NB
NS
PS
PB
NB
NB
NB
NS
Z
NB
NB
NS
Z
PS
NB
NS
Z
PS
PB
NS
Z
PS
PB
PB
Z
PS
PB
PB
PB
C. Defuzzification
After evaluating the rule(s), the system comes up with a
certain output frequency (F) which is then fed to DSP
controller. The F tells the inverter how much (variable)
frequency it has to produce for required heating. Figure 6
shows the output F for two inputs T and T/t.
10
Proceedings
Figure 9 shows the successful generation of SPWM wave on

CCS graph.
Figure 6 simulation output for given input sample values
IV. SPWM GENERATION

SPWM pulses are constant amplitude pulses with different
duty cycles for each period. For SPWM generation reference
signal is compared with carrier (triangular wave). A sinusoidal
waveform signal is used as a reference signal to shape the
output (AC voltage) close to sinusoidal.
Figure 7 shows the comparator circuit for SPWM
generation.
U1
0V
V2
VOFF = 0v
VAMPL = 4v
FREQ = 1khz
R1
OUT
1k
5.000V
OPAMP
0V
V8
V1 = 5v
V2 = -5v
TD =
TF = .00002775
TR = 0.00002775
PW = 5.55e-7 0
PER = 5.55e-5
Figure 9 SPWM generation on CCS graph
V. TEMPERATURE SENSOR INTERFACING

Temperature detectors are necessary element in order to
carry out the temperature controlling. The output of sensor is
connected to DAQ (NI-PXI-6229) to link the sensor with the
software environment.
This output of DAQ assistant is multiplied with sensor
sensitivity. The output of multiplier block is fed to formula
block and collector block. The formula block is converting the
analog signal into oCentigrade and the collector block
produces the mean of collected readings. The collector output
is fed to signal conditioning block. The readings finally
coming out are being written to measurement file block. The
whole circuit is enclosed in a while loop to carry out the
reading process continuously as shown in Figure 10.
Figure 7 Comparator for SPWM generation
For real time frequency variation of SPWM signals, variable

frequency sine wave is created. For sampling frequency of
8KHz, frequency of generated sine wave, Equation 2 is used,
Where n represents number of points inputted for sine wave

generation.
With the change in value of n, the frequency of sinusoidal
wave gets changed. Figure 8 shows the output behaviour of
Variable frequency sine wave generated.
Figure 10 VI window for Temperature Sensor Interfacing

Circuit
The temperature is forced to bring around 70 to 75oC. The

change in temperature is noted down in the write to
measurement file. This data is plotted and the rise and fall
response of temperature is modelled as shown in Figure 11
and Figure 12 below:
Figure 8 Output of Variable frequency sine wave
SPWM signals are generated by comparing triangular wave

(1KHz) with variable frequency sinusoidal wave. The
amplitude modulation ratio is set to 0.9 and frequency
modulation ratio varies as frequency of reference signal
varies, determined by Equation 3;
Figure 11 sensor response when temperature rises
11
Proceedings
http://www.datasheetcatalog.com/datasheets_pdf/I/R/F/8/IRF840.sht
ml
[4] HV Floating MOS-Gate Driver ICs, Application Note AN-978,
http://www.irf.com/techincal-info/appnotes/an-978.pdf
[5] Zememe Walle Mekonnen, Digital Signal Processing
Applications using C6713 DSK, project work
Figure 12 sensor response when temperature falls
[6] S. Zinn, S.L Semiatin, Elements of Induction Heating, Design,

Control and application
VI. INDUCTION HEATING

Induction heating has replaced the traditional furnace
methods because of its efficiency, unpolluted and fast heating
process. Electrically conductive materials are used in
induction heating for heating purposes and this requires high
frequency electricity. An IH system requires a source of
alternating current, an induction coil, and the work piece to be
heated. A magnetic field is generated in the coil due to the
alternating current passing through the coil. The AC is
supplied by the inverter. Work piece placed within the coil
will experienced the magnetic field due to which eddy
currents are induced in the work piece that cause non-contact
type of heating between work piece and the induction coil.
Copper tube is used to make induction coil. The tube is
hollow inside with coil diameter about 3 inches and internal
diameter of tube is about 1cm.the coil is given 8 turns. The
impedance of coil depends on cross-sectional area, length and
the number of turns so to increase the coil impedance.
The impedance matching circuit is designed such that it
converts high volt/low current (coming out from inverter) to
low volt/high current (driven requirement of the load). The
choice is made for impedance matching in order to match the
output parameters of inverter with the input parameter of coil.
Parallel capacitor bank is to be connected between the coil and
inverter.
VII.
CONCLUSION
The FLC for induction heating has been presented in this

paper. The integrated controller, implemented on DSKC6713
takes the set temperature (required temperature of load) input
from user, reads actual temperature of load and calculates
error and rate of error, then on basis of Fuzzy Logic, it takes
decision and determines the amount of F that needs to be
added/subtracted. The FLC then feeds this F to/from
operating frequency of variable frequency inverter.
The FL is being implemented in international industries but
is new for Pakistani industries. However, Pakistan industries
are doing their controlling on fuzzy logic but it is a joint
venture of PID and FUZZY. This idea, however, introduces
Fuzzy Temperature Controller as a standalone product.
BIBLIOGRAPHY
[1] Chin-Hsing Cheng, Design of Fuzzy Controller for Induction
Heating Using DSP, 5th IEEE Conference on Industrial Electronics
and Applications, 2010
[2] Yunseop Kim, Fuzzy Logic Temperature Controller, Physics
344 project, 2001
[3] Datasheet IRF840:
12
Proceedings
Stability and Control Solution to Quad-copters

Q. Ejaz Ur Rehman1, S.Akhtar1, A.Saleem1
Department of Avionics Engineering1
National University of Sciences and Technology
Islamabad, Pakistan
qejaz@cae.nust.edu.pk
Suhail@cae.nust.edu.pk
ammar.saleem@cae.nust.edu.pk
Abstract A Quad-copter is a structurally simple and
dynamically complex rotorcraft, lifted and propelled
by four rotors. It has very small size and is highly
maneuverable as compared to conventional
helicopters. In this paper, a method to achieve control
and stability of a quad-copter is presented.
Computational
tools
employed
are
MATLAB/Simulink, Catia, LabVIEW, ANSYS
and Arduino IDE. A Mathematical model of quadcopter dynamics is developed using set of derived
nonlinear equations accompanied by control theory.
This nonlinear mathematical model is linearized in
Matlab and LabVIEW. Linear equations are used to
design Linear Quadratic Regulator (LQR) controller.
Microcontroller and sensor used are Arduino Mega
2560 and 6 DoF1 IMU2. Stability of quad-copter is
validated through experiments and simulations.
aerial photography, Bomb search and disposal, vision

based pose estimation and Fertilizer/ Pesticide sprayer
etc.
Unlike most helicopters, Quad-copters use two sets of
indistinguishable static level propellers (two
clockwise (CW) and two counter-clockwise (CCW))
which are set in an X or + (plus) configuration with X
being the preferred configuration as shown in figure 1.
These use deviation of RPM to control thrust and
torque. Roll, Pitch and Yaw of quad-copter is achieved
by altering the rotation rate of one or more rotor discs,
thereby changing its torque load and thrust/lift
characteristics.
Keywords: Quad-copter, Stability, Controllability,

LQR, Mathematical Modeling, IMU
I.
INTRODUCTION
In recent years, the aeronautical industry has shown a

growing interest in UAVs (Unmanned
Aerial
Vehicles). UAVs are growing in popularity in fields of
medicine, engineering, civil and most importantly,
military and security. The reduced cost, absence of a
trained pilot and small compact size make them viable
options for tasks that include inhospitable terrain and
remote regions.
Figure 1: Rotors 1 and 3 rotate in one direction, while

rotors 2 and 4 rotate in the opposite direction,
controlling opposing torques for controllability and
stability.
The dynamics linked to employing four rotors
mounted on edges of a square shape create a highly
unstable platform that can only be controlled by
embedding complex algorithms onboard.
Quad-copters are the conventional remote control

airborne vehicles with four rotors placed at equivalent
distance from center of gravity. Quad-copter is
elevated and driven by these four rotors only. Quadcopters are structurally simple and unstable unmanned
air vehicles (UAVs). Due to its simple structure it is
popular UAV nowadays, being used for surveillance,
1
Due to the dynamically unstable nature of rotors,

complex control mechanisms are required for a
sustained flight.
Degree of Freedom
Inertial Measurement Unit
13
In this paper, a method to achieve controllability and

stability of quad-copter at certain height is achieved
such that it is stationery with respect to the earth frame
of reference at certain height. Simulation platform
used are MATLAB and LabVIEW, while detail
study of quad-copter and propeller is conducted in
ANSYS-FLUENT. CAD models are modelled in
CATIA software. The algorithm is written by
manipulating the non-linear differential equations with
control theory. The algorithm written is verified by
visualizing results, animations and virtual reality
model to completely study the quad-copter behaviour
and response to inputs.
Two coordinate systems are considered in Figure 4

[3]:
The algorithm is translated into equivalent C language

for Hardware testing. Microcontroller and sensor used
are Arduino Mega 2560 and 6-DoF Razor IMU only.
II.
Proceedings
The inertial frame (E-frame)

The body-fixed frame (B-frame)
Figure 3: Quad-copter Frame of References

These are related through three rotations:
EQUATION OF MOTION
The governing equations for the control of quad-copter

are derived in this section. First of all, translational and
rotational dynamics of quadrotor are explained
followed by simplifications. Bold symbols are used to
denote three-dimensional vectors, while non-boldface
symbol are used for scalars in the paper.
Roll: Rotation of around the x-axis;

Pitch: Rotation of around the y-axis;
Yaw: Rotation of around the z-axis.
The following assumptions were made in this

approach:
The body-fixed frame origin and center of

mass (COM) of the body of the vehicle are
coincident.
The axes of the B-frame coincide with the
body principal axes of inertia.
Figure 2: (A) Dynamic model of a quad-copter with

four propellers in Earth frame of reference (B)
Propeller i producing fi thrust with rpm in zdirection
A. Dynamics
A quad-copter is a UAV having four rotors and a mass
m. The forces which act on a quadrotor are its weight
and the thrust f produced by four propellers in body
fixed direction z = (0, 0, 1). Similarly, four torques acts
on each propeller and a total drag torque acts on quadcopter body. The rotation of the body fixed frame with
respect to some inertial frame is described by the
rotation matrix R which is discussed in detail.
Figure 4: Quad-copter configuration frame

system
Given equation describes the relation of Rotation
matrix with roll, pitch, yaw and quad-copter position
in earth frame.
R (,, ) R(x,) R(y,) R (z, )
(1)
The main equation governing the quad-copter

dynamics is:
14
m = 4i=1 fi + m
Proceedings
proportionality constant of k and sign given by the

sense of rotation).
(2)
Where d= (d1, d2, d3) are position vectors in inertial

frame of reference and g = (0, 0, -9.81) is gravitational
constant.
i = (1)i+1 fi
And thrust is related to rotor rpm as
As quad-copter is a rigid body with four propellers.

The body inertia is expressed as diagonal matrix
B
Ixx
IB = [ 0
0
0
B
Iyy
0
0
0]
fi = f 2i
(3)
B
T
T )qr
P
Ixx
p = f (w22 w42 )l (Izz
Ixx
Izz
q(w1 +
w2 + w3 + w4 ),
( 11 )
B
Izz
0
P
Iyy
0
0
0]
B
T
T )pr
P
Ixx
q = f (w32 w12 )l + (Izz
Ixx
+ Izz
p(w1 +
w2 + w3 + w4 ),
( 12 )
B
Izz
r = r + f (w12 w22 + w32 w42 )
(4)
III.
P
Izz
The angular velocity of a body can be governed using

the differential equation: [2]
res = B + 4i=1 P + B (LB +
4i=1 I P B + 4i=1 LP )
(5)
IV.
Where res is the resultant torque acting on quadcopter body. B is the body torque
(6)
(f2 f4 )l + dx
(f3 f1 )l + dy
]
1 + 2 + 3 + 4 + dz
EXPERIMENTAL PLATFORM
CAD models of Quad-copter and equivalent propeller

were modelled in CATIA and were exported to
ANSYS-FLUENT where surface meshing and
computational fluid dynamics (CFD) was done to
study aerodynamic design such that fluid was passed
on to the quad-copter and propeller at different speeds
and direction to check its serviceability.
(7)
B. Simplification of Assumptions
The effect of all the moments acting upon the body is
denoted by res on the right hand side of equation (5).
These include moments due to propeller forces and
torques due to motors. The propeller forces are
assumed to act through the center of each propeller. It
is assumed that the center of propeller is at a horizontal
distance l from the body center of mass.
res = [
CONTROLLABILITY
Computational Programming softwares employed

were MATLAB / SIMULINK, LabVIEW,
ANSYS, Arduino IDE and CATIA.
And P is the torque produced by propeller.

P
P = (0, 0, Izz
i )
( 13 )
These non-linear equations are linearized to compute

state space matrices A, B, C and D. An LQR controller
was designed with the cost value of 0.5 s2 rad-2 on the
angular rates, 10 on the deviation from the primary
axis, 0 on the extended motor states and 0.75 N-2 on
the inputs.
P
P
Due to symmetry of propeller Ixx
= Iyy
.
B
B
B
B = (Ixx
p , Iyy
q , Izz
r )
( 10 )
Solving the above equation from 3 to 8 results in [2]
The propeller is a symmetric body with respect to its

axis of rotation and can be considered as disc for
simplification. The propeller inertia is given by:
P
Ixx
IP = [ 0
0
(9)
The propeller produced vibrations which were verified

from CFD analysis. These vibrations were catered
using prop-balancer system.
(8)
With d = (dx , dy , dz ) is the drag torque. It has

been observed that propellers reaction torque has a
linear relation with the thrust force (with
15
V.
Proceedings
MODELLING OF PROPELLER
A. Actuator Disk Theory

A propeller can be represented as a single disk
operating in a stream tube. As the flow passes through
the Disk its velocity decreases while pressure
increases. The disk is infinitely thin but has an area.
Here propeller acts like it is made up of infinite blades.
This disk produces pressure jump across it which is
equal to thrust per unit area of disk.
Figure 7: Thin enclosed in disk
Figure 5: Actuator Disk in flow stream
B. Geometry of Fan
In order to model the propeller a thin circular
surface of area equal to swept area of propeller was
created around the hub. This thin surface was enclosed
in a disk shaped volume such that the diameter of disk
and thin surface was same. Two fluids, fluid fan 1 and
fluid fan 2 were defined on both sides of thin in
between thin and disk surface.
Figure 8: Volume Mesh with Quality
Figure 9: Vectors through thin

The inertia of the quad-copter was measured by
suspending the quad-copter about 3 different axis and
measuring its period of oscillation. The inertia is given
by:
Figure 6: Thin with fluid fan
I = (2T)2
( 14 )
Where k is the torsion constant and T is the period of

oscillation of quad-copter with reference to
16
equilibrium
position.
P
P
symmetry Ixx
= Iyy
.
Due
to
Proceedings
quad-copter
The propeller inertia was approximately measured by

considering propeller as disc and motor as cylinder. As
P
P
propellers are fixed in zth direction so, Ixx
= Iyy
=0
The quad-copters mass was determined to be 1 Kg.
The distance from the center of gravity of the quadcopter to center of propeller was 0.24m. The propeller
reaction torque and force constant were estimated as
= 0.214272
Nm
N
and f = 1.80899e 5
Ns2
rad2
Figure 10: Voice Command Panel in LabVIEW for

quad-copter
.
VII.
VI.
VOICE CONTROL PANEL
Mathematical Model
A mathematical model was developed in LabVIEW

and MATLAB using nonlinear equations discussed
in section II. Implemented nonlinear equations were
linearized to design LQR controller for feedback
control. Implemented mathematical model, results and
animation is shown:
A simulation based voice control panel was also

developed in LabVIEW to control quad-copter in 6DoF which is shown in Figure 8.
It operates on the certain commands and performs
tasks as per the commands which are embedded in the
algorithm.
Figure 11: Non-Linear Mathematical Model
17
Proceedings
Figure 14: Animation GUI of Hover Model

An animation GUI has also been implemented which
shows the behavior of quad-copter and also display the
trajectory made by quad-copter to reach height of 10
ft.
Figure 12: Simulation Results
From figure 12 it can be seen p, q, r (angular rates
(degs/s) about x, y and z axis, respectively) showed
variation at the start of quad-copter but came to
equilibrium position at t=2, 5, 3 sec respectively.
Subsequently roll, pitch and yaw angle also settled
after 5 sec of takeoff. While linear velocities U and W
about x and z axis rises continuously and linear
velocity V about y-axis comes to equilibrium position
after 10 sec which depicts the quad-copter has gained
height of 10 ft. above ground as justified from X, Y, Z
in the figure 10.
Figure 15: Virtual Reality Quad-copter Model

A quad-copter model has also been made using
Matlab virtual reality toolbox in order to make the
simulations near to reality. This virtual model take
roll, pitch, yaw and rotor forces as input and depicts
the result of the mathematical model. It was stabilizing
itself after takeoff, initially showing some small
vibrations which can also be visualized from figure 10.
VIII.
Figure 13: Throttle and RPM relation with time
CONCLUSION
Quad-copter is a highly unstable UAV, but due to its

high maneuverability, it is highly desired for field
works. Utilizing nonlinear dynamic equations
accompanied with control theory can bring quadcopter to life. These algorithms sufficiently reduce the
need of pilot and can be used to build cheap UAVs
which can reduce cost to a great extent. Quad-copter
are the future robots in this field in particular as well
as in other fields in general, being applied alongside
Figure 13 shows the variation of produced thrust from

throttle given at certain time t for each rotor.
18
people which will help in making tasks easier and

more efficient. Developed mathematical model was
successfully implemented on hardware.
IX.
ACKNOWLEDGMENT
This work has been made possible by the help of my

co-authors which include my project advisor and coadvisor. This text has been rectified and proof read by
Undergraduate students M. Moghees Shahid and Ali
Mahmood. They are destined for great things.
This project has been sanctioned by College of
Aeronautical Engineering, NUST.
REFERENCES
[1]
T. Luukkonen, "Modelling and control of

quad-copter," Independent research project
in applied mathematics, Espoo, 2011.
[2]
M. W. Mueller and R. D'Andrea, "Stability

and control of a quadrocopter despite the
complete loss of one, two, or three
propellers," in Robotics and Automation
(ICRA), 2014 IEEE International Conference
on, 2014, pp. 45-52.
[3]
I. Gaponov and A. Razinkova, "Quad-copter

design
and
implementation
as
a
multidisciplinary engineering course," in
Teaching, Assessment and Learning for
Engineering
(TALE),
2012
IEEE
International Conference on, 2012, pp. H2B16-H2B-19.
[4]
D. Hanafi, M. Qetkeaw, R. Ghazali, M. Than,

W. Utomo and R. Omar, 'Simple GUI
Wireless Controller of Quad-copter',
International Journal of Communications,
Network and System Sciences, vol. 06, no.
01, pp. 52-59, 2013
[5]
Y. Cooper, R. Ganesh Ram, V. Kalaichelvi

and V. Bhatia, 'Stabilization and Control of
an Autonomous Quad-copter', AMM, vol.
666, pp. 161-165, 2014.
19
Proceedings
Proceedings
Detection of Fire Hotspots dealt by Emergency

First Responders in Rawalpindi using GIS
applications
Amna Butt and Sheikh Saeed Ahmad
Department of Environmental Sciences
Fatima Jinnah Women University
Rawalpindi, Pakistan
ambutt91@yahoo.com; drsaeed@fjwu.edu.pk
The study area of the present research was Rawalpindi city

(Fig. 1). The citys administrative boundaries consist of two
tehsils namely Rawal and Potohar tehsil. Currently five service
stations and two key points of Rescue 1122 (otherwise known
as Punjab emergency service) are providing emergency
response services (including fire brigade service) in different
areas of the city. The resources currently available to them for
providing Fire brigade service include 9 fire vehicles, 14
ambulances and personnel of 400 trained rescue providers.
However, no prior study has been conducted in the city focused
on surveillance of fire emergency response. Furthermore, the
existing management strategy for improvement of fire
emergency response is not very effective and no thought has
been given to incorporating GIS expertise in the department for
this purpose.
Abstract During the past, need of efficient Emergency First

Response (EFR) for Rawalpindi, along with all the major cities of
Pakistan has increased tremendously. Therefore, there is a need
to develop an effective management strategy to improve the first
response services. Present study focused on the identification of
past and current service locations for fire incidences and
mapping these locations for hotspot identification. The incidence
data for past five years (2009-13) was collected and Hotspot and
Spatial Autocorrelation analyses were performed on the data to
detect the fire hotspots and their clustering patterns in the city.
The results revealed a slight shift in fire hotspots in 5 years and
also in the clustering pattern which changed from significantly
clustered (2009) to randomly distributed (2013). Hotspot and
spatial distribution maps were generated to indicate the fire
hotspots in the city. These maps can be helpful to prevent the
future incidents by allocating more service stations focusing these
areas for fire mitigation.
Index Terms Hotspot Analysis, Fire, Kernel Density,
Emergency First Response (EFR), Spatial Autocorrelation
Analysis
I. INTRODUCTION
Coping with fire, caused either by natural or anthropogenic
factors is one of the challenges faced by the modern societies
[1]. Analyzing the city fire risk is therefore highly significant
for development of effective urban fire protection plan and
regulations and facilitates the coordinated development of
social economy [2].
Application of geostatistical tools of GIS can play a
significant role in improvement of local fire emergency
response [3] primarily by facilitating the visualization and
interpretation of nature and previously observed patterns of
such accidents [4][5]. Generating different fire risk maps on the
basis of geostatistical analysis is also imperative to develop
strategies focused on alleviating the future risk [6].
Numerous approaches based on GIS have been developed
and used over the past to provide geostatistical surveillance of
the precedent emergency patterns for development of several
models for fast and apt response delivery [7][8][9][10][11].
A. Study Area
Fig. 1. Study Area map: Rawalpindi City
20
Therefore, the primary objective of the present research is

to provide a GIS based surveillance system for the Fire
incidents of the past in order to determine the recurrent service
locations for focusing the future response. The study can
therefore be considered as a baseline for future improvement in
the quality and efficiency of fire emergency responses in
Rawalpindi city.
Proceedings
threshold limits ranging from 500-2000 meters. The range used

for determination of correlation was -1 to 1 and Z-score value
was calculated to assess the statistical significance of the
observed clustering (based on correlation) for each year. The
highest correlation values were then recorded for each year and
subsequently were employed for hotspot identification.
2) Hotspot Analysis: Getis-Ord Gi*
Fire hotspots were identified based on the Getis-Ord Gi*
statistics. For this purpose, the conceptualization of spatial
relationship among different datasets was done by opting
Zone of indifference. The threshold limit was set on the basis
of spatial autocorrelation outcomes for each year (exhibiting
highest Z-score value). Thereafter, the identified hotspots for
Fire cases were interpolated using Inverse Distance
Weighted or IDW for better visualization of results and
hotspot maps for each year were generated.
II. MATERIALS AND METHODS

The research methodology that was applied to obtain the
results primarily consisted of four main steps incorporating
data collection, processing, analyzing and visualizing the
results (Fig.1).
III. RESULTS
The emergency callout data on building fires, bursting of
gas pipes, cylinder blasts, and gas leakages was cataloged in
the category of Fire emergencies. Different geostatistical
analyses were then performed to determine the pattern of
emergency cases for the study duration. The reported incidence
of FE cases for the time period of 2009-2013 were 671. Out of
which, 583 (86.9%) were males and 88 (13.1%) females. 15%
of the total fire incidents (102 cases) were reported in 2009 and
37% (247 cases), the highest incidence, in 2010. After 2010 the
incidence rate declined progressively from 24% (163 cases) in
2011 to 10% (66 cases) in 2012 and subsequently rose to 14%
(93 cases) in 2013 (Fig. 3).
Fig. 2. Flowchart of main steps followed in methodology
B. Data Collection and Processing

The data collected for the purpose of this study was divided
into two categories namely Primary and Secondary data. The
main data obtained for the purpose of the study was secondary
data acquired from the headquarters of Rescue 1122,
Rawalpindi in form of caller and victim directory. This data
was collected on unit level i.e. data from all the emergency
units of 1122 at work in Rawalpindi city was acquired,
compiled and then processed for segregation of Fire cases.
Primary data was then collected on the basis of segregated fire
incidents. This data comprised the GPS readings of incident
locations obtained via handheld Oregon 650 GPS for the
reported fire cases.
Both primary and secondary data was processed in
Microsoft excel and then loaded to ArcGIS 10.2 for further
processing and analyzing.
Fig. 3. Percentage Contribution of Fire Incidents in Rawalpindi during 20092013
C. Geostatistical Analysis of data

The data was geostatistically analyzed in ArcGIS 10.2
environment for determination of spatial clustering and
identification of hotspots. The geostatistical analysis performed
for this purpose included Global Morans I test (Spatial
autocorrelation analysis) and Hotspot Analysis (Getis-Ord
Gi*).
1) Global Morans I statistics
Global Morans I statistic gives an indication of any
existing correlation among spatial observations and delineates
the characteristics of the global pattern. The pattern maybe
random, dispersed or clustered depending on the spatial
association present in the data [12][13]. For the purpose of
present study spatial autocorrelation among the fire incidents
was calculated on yearly basis by employing different
The possible spatial autocorrelation of Fire cases estimated

by Morans I statistics revealed significant spatial clustering for
the years 2009 and 2011, whereas 2010 showed mild
clustering. 2012 and 2013 on the other hand showed random
patterns. Morans I and G-statistic values (Z-score) for all the
years, given in Table 1 disclosed that 2009 had highest (20.01)
while 2013 had lowest (1.10) Z-score.
21
Proceedings
Table 1. Global spatial autocorrelation statistics of Fire emergencies for 20092013

Year
Moran's I
Z Score
P Value
Pattern
2009
2010
2011
2012
2013
0.77574
0.01719
0.17427
0.39961
0.02737
20.0117
2.09997
4.3162
1.51482
1.10479
0.001000
0.035732
0.000016
0.129818
0.269251
Clustered
Clustered
Clustered
Random
Random
Local Gi* (d) statistic employed to identify the hotspots for

Fire incidents in Rawalpindi during 2009-2013 categorized the
Z-score outcomes at 5% significance level as either clusters or
non-clusters. The identified Fire hotspots covered both urban
and rural areas of Rawalpindi city for all the years. The specific
hotspot locations for each year are tabulated in Table 2
however.
Table 2. Identified Fire Hotspots for the years 2009-2013
Year
Identified Hotspot Locations
Asghar Mall Chowk, Banni Chowk, Chaklala Scheme 3, Chandni
2009
Chowk, Islamabad Highway, Khanna Pull, Link Road, Murree Road,
Muslim Town, Naz Cinema Chowk, Rehmanabad, Sadiqabad
Chowk, Transformer Chowk and Waris Khan Stop
A.R.I.D University road, Band Khanna Road, Bilal Hospital road,
2010
Dhok Kashmiriyan, Double Road, Faizabad, Ghosia Chowk, Iqbal
Town, IJP road, Kattariyan, Khayaban-e-Sirsyed, Kurri road, Rabi
Center, Saidpur road Satellite Town, Sixth road, Shamsabad and
Sohan Pull
Dhok Mustakeem, Choor Chowk, Golra Morr, Misriyal road,
2011
Peerwadhai Morr, Qasimabad, Seham road and Westridge
Askari 11, Faisal Colony, Jhanda Chichi, Military Hospital Road,
2012
Peshawar Morr, Pindora Chungi and Shamsabad Stop
Adyala road, Chaklala Scheme 3, Committee Chowk, Khanna Pull,
2013
Link road, Raja Bazar, and Rawal road
Figure 4 revealed that Fire hotspots for 2009 and 2010 were
mostly contained in the Northern region of Rawal Tehsil and
shifted towards North-West in 2011. However, during 2012
and 2013 not only the number of hotspots reduced significantly
but the spatial distribution pattern also became random.
Fig. 4. Mapping of Fire hotspots using Getis-Ord G* statistics during 20092013
Spatial distribution of Fire cases in hotspot locations was

also visualized by creating spatial distribution maps (Fig. 5).
The highest incidence (based on the number of cases per
locale) was observed in 2010 while the lowest was observed in
2012. The distribution pattern revealed that the highest
incidence was mainly in urban areas of Rawal tehsil and NorthEast portion of Potohar tehsil where the reported cases per
locality per year were as high as 10.
22
Proceedings
organizations and attempting to solve the matter themselves.

However the accounted figures do not represent the total
number of cases observed in the city, just the incidence that
was dealt by Rescue 1122. Other organizations at work in the
city for fire control and management include Rawalpindi Fire
Brigade Center and Qureshi Fire Control Services.
Akhter [16] explicated the reasons behind the lack of
confidence among public. The study reasoned that there is
disparity in implementation of fire safety standards in the city
as well as very little coordination among different departments
such as traffic, police and fire fighting units. This lack of
coordination along with unavailability of Incident Command
System (ICS) often translates to poor emergency response
despite good skills and training. Dawn [17] conversely reported
that local fire brigade services lack in performance due to
insufficient professional training, availability of resources,
planning and research (both pre and post fire) and nonexistence
of any fire services act for the city. All these factors, along with
lack of awareness among general public regarding fire fighting
profession has a negative impact on fire emergency response.
Hence, for improvement of future fire emergency response,
it is need of hour to understand the previous and existing
patterns, risk factors and causative agents; and ensure effective
enforcement of building and fire safety laws [18].
V. CONCLUSION AND RECOMMENDATIONS
Present study focused on providing a geostatistical
surveillance approach for ensuring future fire safety by
improving the response quality and apt resource allocation for
high risk areas. Based on the outcomes of the research, it is
concluded that there is both spatial and temporal variation in
the occurrence of Fire incidents in the study area. Most of the
Fire hotspots however were located in the Northern portion of
the study area incorporating both urban and rural areas of the
study indicating the need to shift the focus of fire service in this
region of the study area. The study further concludes that there
is a need of incorporation of GIS based surveillance system in
the rescue department to direct the response from the service
stations in a timely manner.
Therefore, the study recommends generating awareness
among people regarding fire hazard and the factors associated
with it, incorporation of GIS expertise in emergency
departments, promotion of GPS enabled cell phones in dispatch
units and fire vehicles, high level of collaboration among
different departments working in the city for fire services
provision to avoid service duplication and publication of GIS
based maps and models designed for response improvement.
Fig. 5. Spatial Distribution pattern of Fire incidents in Rawalpindi for study

duration
IV. DISCUSSION
In order to facilitate the efficient management of fire
emergencies in an area, improvement of existing response
systems is of high significance [14]. This can be ensured by the
providing surveillance for past occurrences and understanding
of recurring patterns.
Significant Fire hotspots were manifested in both urban and
rural areas of the city and were mainly contained in Northern
portion of the study area. These were the areas having high rise
buildings, gas stations, commercial areas, suburban areas,
highways and residential areas (with heavy load shedding of
gas). As these hotspots were estimated not only for household,
commercial and secondary fires but vehicular fires as well,
various roads were also identified as hotspots. Mostly the Fire
emergencies were observed on the roads that are used by heavy
vehicles as they are more prone to overturning and catching
fire. Corcoran et al. [11] also analyzed the spatial patterns of
fire by employing GIS and obtained similar results. Rao [15]
also reported similar findings and additionally said that the
reason for Fire incidents in the city is exposed and jumbled
cable wires made of substandard material.
However, the results of the study indicated a significant
decline in the intensity of cases for the duration of the study
(Fig. 3). This declining incidence gave an account of lack of
confidence in general public to refer to firefighting
ACKNOWLEDGMENT
We are endepted to the department of Rescue 1122,
Rawalpindi for providing the data regarding fire emergencies
and cooperating with us throughout this research.
REFERENCES
[1] M. I. Channa, and K. M. Ahmed, Emergency Response
Communications and Associated Security Challenges, Int. J.
Net. Sec. and its Appli., vol. 2 (4), pp. 179, 2012.
23
Proceedings
[10] A. Spoerri, M. Egger, and E.V. Elm, Mortality from road

traffic accidents in Switzerland: Longitudinal and spatial
analyses, Accid. Anal. and Prev., vol. 43, pp. 40-48, 2011.
[11] J. Corcoran, G. Higgs, C. Brunsdon, A. Ware, and P. Norman,
The use of spatial analytical techniques to explore patterns of
fire incidence: A South Wales case study, Comp., Env. and
Urban Sys., vol. 31, pp. 623-647, 2007.
[12] B.N. Boots, and A. Getis, Point Pattern Analysis Newbury
Park. Newbury Park, CA, USA: Sage Publications, 1998.
[13] L. Fang, L. Yan, S. Liang, S. J. D. Vlas, D. Feng, X. Han, W.
Zhao, B. Xu, L. Bian, H. Yang, P. Gong, J. H. Richardus, and
W. Cao, Spatial analysis of hemorrhagic fever with renal
syndrome in China, BMC Infect. Dis., vol. 6, pp. 77-88, 2006.
[14] S.R. Morgan, A. M. Chang, M. Alqatari, and J. M. Pines,
NonEmergency Department Interventions to Reduce ED
Utilization: A Systematic Review, Acad. Emer. Med., vol. 20,
pp. 969-985, 2013.
[15] S. Rao, Rescue 1122 management plan finalized, The
Nation. Retrieved from http://nation.com.pk/karachi/06-Jan2010/Rescue-1122-management-plan-finalised, 2010.
[16] S. Akhter, Firefighters view on Improving Fire Emergency
Response: A Case Study of Rawalpindi, Int. J. Hum. and Soc.
Sci., vol. 4(1), pp. 143-149, 2014.
[17] Dawn, Pindi fire brigade squad runs out of steam, Dawn.
Retrieved from http://www.dawn.com/news/90148/rawalpindipindi-fire-brigade-squad-runs-out-of-steam, 2003.
[18] Pakistan Observer, 1122 Rawalpindi Rescued 1883 Emergency
Victims,
Pakistan
Observer.
Retrieved
from
http://pakobserver.net/detailnews.asp?id=251475, 2014.
[2] W. Aiyou, S. Shiliang, L. Runqiu, T. Deming, and T. Xiafang,

City Fire Risk Analysis based on Coupling Fault Tree Method
and Triangle Fuzzy Theory, Proc. Engg., vol. 84, pp. 204-212,
2014.
[3] Environmental Systems Research Institute (ESRI), Improving
Emergency Planning and Response with Geographic
Information Systems, Redlands, New York: ESRI. Retrieved
from http://www.esri.com/library/whitepapers/pdfs/emergencyplanning-response.pdf, 2005.
[4] S. Erdogan, I. Yilmaz, T. Baybura, and M. Gullu,
Geographical information systems aided traffic accident
system case study: city of Afyonkarahisar, Accid. Anal. and
Prev., vol. 40, pp. 174-181, 2008.
[5] M. Kwan, and J. Lee, Emergency response after 9/11: the
potential of real-time 3D GIS for quick emergency response in
micro-spatial environments, Comp., Env. and Urban Sys., vol.
29, pp. 93-113, 2005.
[6] C. Yan-yan, L. Dong, and Z. Hui, Multi-factor Risk Analysis
in a Building Fire by Two Step Cluster, Proc. Engg., vol. 11,
pp. 658-665, 2011.
[7] M. H. Hussain, M. P. Ward, M. Body, A. Al-Rawahi, A. A.
Wadir, S. Al-Habsi, M. Saqib, M. S. Ahmed, and M. G.
Almaawali, Spatio-temporal pattern of sylvatic rabies in the
Sultanate of Oman, 20062010, Prev. Vet. Med., vol. 110, pp.
281-289, 2013.
[8] T. Ruya, M. Ning, L. Qianqian, and L. Yijun, The Evolution
and Application of Network Analysis Methods, IEEE Int.
Conf. on Sys., Man, and Cyber., pp.2197-2201, 2013, DOI
10.1109/SMC.2013.376.
[9] D. Dai, Identifying clusters and risk factors of injuries in
pedestrianvehicle crashes in a GIS environment, J. Trans.
Geo., vol. 24, pp. 206-214, 2012.
24
Proceedings
Prospects of Airborne Wind Energy Systems

in Pakistan
Z. H. Khan
Dept. of Electrical Engineering
Riphah International University
Islamabad, Pakistan
zeeshan.hameed@riphah.edu.pk
Sohaib Khan
Dept. of Aerospace Engineering
Islamabad, Pakistan
sohaibkham786@gmail.com
Arsalan Khan
Dept. of Control and Simulation
CESAT
Islamabad, Pakistan
arsalan1@mail.nwpu.edu.cn
This limitation greatly affects the utilization of WECS as an

effective resource of energy as usually far off areas near seashores or mountains fulfills the required energy density criteria
for feasible investments. Also, long way cabling for
transmission of this energy to grid also results in additional
cost and line losses [3]. Social activists and NGOs raise voices
and show their grievances against environmental impact of
these wind farms due to ever increasing bird accidents as of
result of collision with fast moving blade tips.
Abstract High altitude wind energy is considered as a high

efficiency, low cost solution to sustainable energy solutions
worldwide today due to highly flexible and adaptive designs of
powered kites. Pakistan is facing energy crisis since years due to
inefficient electrical transmission network, increased demand of
electricity, lack of resource planning and increased cost for
furnace oil which can be used to generate electricity. As a clean
energy source, conventional wind energy is a preferable choice
but it has few disadvantages due to large initial investments
involved and dangers to environment due to rotating machinery
which can result in bird hit and on the other hand, generating
acoustic noise which affect populations of people living nearby.
In response to these issues, high altitude energy is found to be
free from many such problems as found with conventional wind
energy systems. It is found suitable specially for addressing
essential energy needs of off-grid consumers for water pumping,
drilling, refrigeration of vaccines and life-saving medicines and
powering up far-off residential sites e.g. communities living offgrid at Alaskas northern region having minimal solar energy
during the long winter season when energy demand is greatest
for heating purposes. In this paper, we propose a UAV design
which can possibly be used as an airborne wind energy system
for electricity generation.
KeywordsAltitude wind; renewable energy; powered kites;
wind energy conversion system; distributed energy
I.
INTRODUCTION
Energy production for a worlds economy is directly linked

with GDP growth. Renewable energy solutions are a preferable
choice to address the energy demands world-wide due to
environment friendly green energy technology. It has been
estimated that till 2050, 40% of the world energy requirements
will be meet by renewable energy including solar (photo
voltaic (PV) and thermal), wind, bio-mass, bio-fuel etc. [1].
Among various technologies, wind energy is preferred due to
continuous production of electricity (as long as wind is present
as prime mover) as well as high production rate with lesser
requirement of installation area as compared to solar energy
[2]. However, a detailed analysis of wind availability and onsite surveying is mandatory for an optimal production of
energy throughout the year. The conventional wind energy
conversion systems (WECS) are popular only in those areas
where sufficient wind is available i.e. in the range of 5-8 m/s2.
Fig 1. Airborne energy production connected to grid
This paper describes some key design features of an

airborne wind energy system (AWES) e.g. a large kite which is
flown in air while maintaining its connection with a generator
kept at ground level by a suitable rotating mechanism such as
cable drum which is linked via tether. The aerodynamic forces
acting on the kite causes traction in the cable drum which is in
turn converted into electrical power by the generator as shown
in Fig 1. The obvious advantages of shifting the heavy
mechanical components on ground result in simple design and
maximum power optimization. Due to the fact that the flying
kite operates in periodic cycles alternating between traction
phase and reel-in phase of the tether, the electrical power
produced is not continuous rather intermittent which is not
25
suitable for on-grid connection where continuous power is

required. However, a continuous power generation system can
be designed either by using multiple, individually controlled
kites or a battery system for buffering the power generation
across the cycles. This solution is guaranteed to generate
higher power with lower cost as compared to conventional
wind energy conversion system (WECS) due to availability of
faster wind speeds at high altitude.
Proceedings
C. Mobility of the power station

In comparison with WECS, the airborne energy converter
can be moved anywhere. Example include rural areas, Coastal
belt, Desert, etc. However, if flying at higher altitudes, caution
must be paid to operate AWES in no-fly zones (NFZ) to avoid
any danger for civil aviation.
D. Efficiency
The advantage of increasing altitude as in the case of
AWES is to get more capacity due to persistent wind as
compared to turbulence due to buildings and infrastructure at
WECS heights [9].
Fig 2. KGS for generating electricity on-board ship
II. HIGH ALTITUDE WIND ENERGY POTENTIAL

Wind energy is a renewable resource which is not only
cheap but also readily available in most parts of the world.
Wind is discontinuous source of energy and the conventional
WECS provides uneven power supply when connected to the
grid. On the other hand, if altitude is raised, a lot more energy
is available as compared to that blowing at 50-200 m above
ground. It has been estimated that the magnitude of wind
energy above 1000 m is twice as much as found at lower
altitudes. In addition, at an altitude of 800 m above ground,
wind power density is sufficiently available to be used for
altitude energy generation all over the globe [4]. As a rule of
thumb, it has been found that a five-fold cost saving with twice
capacity increase can be achieved in shifting technology from
WECS to AWES [5].
III.
Fig 3. Buoyant Airborne Turbine (BAT) for stand-alone energy generation in

Alaska [10]
IV.
CLASSIFICATION OF AWES
The airborne wind energy systems are classified based on

their designs, grid connectivity and aerodynamics e.g. heavier
than or lighter than air configuration. Some types are indicated
as follows:
1) Flying tethered airplane and kite
In this type of AWES, a tethered kite or airplane flies in
circular or Lissajous shaped orbit to produce electricity via
traction force applied on a generator on ground or ship as
shown in Fig 2. The kite generator system (KGS) is one of the
most commonly used system due to its simplest design and
easy control system [11].
ADVANTAGES OF AWES OVER WECS
A. Impact on Environment
The airborne wind energy systems have fewer effects on
environment as compared to WECS. The fast moving blades of
classical wind mills injures many birds flying in close
proximity. On the other hand, AWES has no dangerous edges
which can kill birds [6].
2) Multiple wing system

This type of altitude wind system has multiple wings to
generate electrical energy. A laddermill is an example of such
system [12]. Multiple wings allow scalability of the concept in
order to generate more energy.
B. Cost comparison
Using AWES instead of WECS, an advantage of more than
90% material savings is guaranteed. This is because large
structures require more and more material to with stand heavy
loads due to wind and the rotating machinery [7]. Many tones
of weight loaded on large erected structures comprises of rotor
blades connected to a hub which drives the generator.
Moreover, the maintenance of WECS is also an issue which
requires access to the system, long down times in case of fault
and danger to the life of technician all adding up the cost
compared to AWES where generator and accessories are
installed on ground [8].
3) Lighter than air systems

These systems are lighter than air and have an on-board
generator to produce electricity as shown in Fig 3. Such
systems are supported by a lighter than air filled balloon to
reach at 100m or above heights where enough wind is
available to drive the generator [13]. Such systems are more
suitable for areas of the world where sunlight is not available
throughout the year.
26
The flight test condition is selected as Mach 0.05 at an

altitude of 200 meters above ground level. Then the angle of
attack () in degrees is varied from 0 to 5 degrees, while Xcg
of 0.9 meter is assumed. The aero-data is plotted in Fig 5 as a
function of angle of attack. It is important to take into account
at least 10% drag increase due to tether connecting the kite
with the ground generator.
4) Rotor hover craft systems

These systems hover in air via rotating propellers.
Quadrotor type hovercrafts are popular which have the ease of
vertical take-off and landing operations [14]. Due to on-board
generators, the airborne system is heavier as compared to the
crosswind kite generator in which has the heavy machinery on
ground.
V.
Proceedings
FLYING WING GENERATOR SYSTEM DESIGN
In order to visualize a practical system, we can use a flying

wing generator (FWG) for our case study. Our goal is to
evaluate optimal design parameters which can be used to
generate maximum wind at different altitudes. By using a
strong flexible tether, energy kites can reach higher altitudes of
100 to 400 meters). This can save up to 90% of the materials of
WECS used conventionally, resulting in reduced per unit
energy cost. As, these systems are more aerodynamic and can
access higher energy density due to stronger winds, each EKS
can generate 50% more energy than their conventional
counterparts [15].
A. Aerodynamics
The wind energy that can be converted to electricity is
given as:
C
2
P
Av w3 C L L
27
CD
Fig 5. Aerodata plots for FlyPG @ Mach = 0.05
(1)
B. Control and Autopilot
Modern design of FWGs incorporates automatic take
off/landing and flight for robust operation under varying
environmental conditions [16]. A coordinated control
mechanism is also devised by few manufacturers where
communication between the Kite/airplane controller and main
controller at the ground station occurs to track the given
trajectory [17].
The relation provides some significant information about

some key design parameters. Equation 1 states that in order to
increase the power, the key parameters to increase include the
wind speed, gliding ration (L/D), the lift coefficient and the
wing surface. As a comparison, energy kites have lower L/D,
price and weight as compared to flying wing. In theory, about
60 kW can be produced per sq. meter of flying kite generator
system. The CFD model is shown in Fig 4.
(a) Side View
(b) Top View
Generally, a multi-objective loop controls the dynamics of

the system in flight. First of all, an optimal track point on
circular or Lissajous trajectory needs to be calculated and
tracked during flight [18]. In most cases a navigation loop
forms the outer one which controls the bearing of the flying
system, while an inner attitude controller controls the roll, pitch
and yaw angle as well as respective rates in order to achieve
the required bearing.
(c) Front View
Fig 4. FlyPG simulation for aerodynamic analysis
In order to analyze the aerodynamic performance of our

benchmark system FlyPG, different flight conditions are used
for Aerodata generation. The aircraft wing is taken as NACAW-4-4410 which has a span of 4 meters, Sref equal to 1.5 m2,
Cbar = 0.38 m, Root chord = 0.35 m, Tip chord = 0.15 m. For
the vertical tail design, NACA-V-4-0010 is selected with Root
chord = 0.25m and Tip Chord = 0.1 m. The horizontal tail
design is NACA-H-4-0010 based with root chord = 0.25 m
and Tip chord = 0.1 m.
Fig 6. A generalized 2-loop control structure for Flying wing/ Energy Kite
As shown in Fig 6, an outer loop controls the bearing

(com) of the flying wing. The error between the commanded
and measured bearing drives the attitude controller in order to
generate roll, pitch and yaw commands for the flying wing
until it reaches the desired navigational coordinates. In this
27
control loop, however, the cross track controller is not

included which is necessary to keep the flying wing on the
desired trajectory by removing the off-track distance [19]. In
practice, stable and robust control is required for fast tracking.
Proceedings
VI.
ENERGY NEEDS IN PAKISTAN
Pakistan is the 6th largest country of the world with a total

estimated population of 175 million people. However, 24%
population lives below the poverty line in the country. As per
the economic survey, it is among those countries of the world
which have lowest per capita energy consumption. The
installed electricity generation capacity is around 22,500 MW.
While the current available power is 17,000 MW, the peak
demand raises up to 22,000 MW in summer resulting in a
shortfall of 5000 MW.
C. Mechanical Subsystem
The mechanical part of a FWG includes wing design after
aero-foil selection. The control actuators (motors) also need
positioning and the embedded controller is designed w.r.t
nominal torque load as well as keeping its maximum limit as
per maximum torque demand. The selection of tether and its
linkage with the ground generator is also important [20].
However, the drag added by the tether length must be added in
the aerodynamic drag analysis for the autonomous controller
design [21].
The economic advisory council of Ministry of Finance

(Govt of Pakistan) foresees an increase in energy demand up to
122.46 M.MTOE in 2022 [27]. In order to satisfy this demand,
at least 12% of this energy is planned to be obtained from
renewable resources in order to relax the huge demand of
furnace oil exports from 30% at present to 20% in 2022 [28].
Table 1 describes the energy demand and forecast which shows
that a total of 122.46 M.MTOE demand is expected in 2022
out of which 14.70 M.MTOE will be obtained from renewable
energy. Looking at the rise in using wind energy as the
renewable source in Pakistan as shown in Fig. 7, it is
emphasized that Altitude wind projects should also be
commenced in the country as a future energy resource
especially for people living in off-grid areas of Pakistan.
D. Electrical Subsystem
The power generation system usually comprises of a
permanent magnet synchronous machine (PMSM) which acts
as a generator [22]. In general, the AWECS can function either
as stand-alone (off-grid) or on-grid. However, an obvious
problem for grid based operation requires an additional cost of
storage system or operating multiple systems at the same time
so that the pumping cycle of one kite is out of phase as
compared to the pumping cycle of the other kites [23]. Thus, at
the instant one kite is relaxing; the other will be supplying
energy to the grid. However, this strategy complicates the
design of the AWECS.
In order to enhance sustainable energy trend, Government

has exempted renewable energy equipment from income
tax/withholding tax and sales tax [29]. Also, import duties have
been waived off for import of such machinery and tools to be
used to develop sustainable energy solutions. Thus, already a
favorable environment is available for foreign investors to
come and finance green energy solutions. Fig. 7 shows a
potential rise of WECS usage as a renewable energy source.
However, the feasibility is only limited to areas where
sufficient wind is present. These areas include the coastal belt
near Karachi (Sindh province), Kati-Bander, Makran and
Jhampir in Balochistan [30, 31].
E. Communication Subsystem
The communication system is required in order to control
the system and for data acquisition. Wireless networks are a
preferred choice in order to reduce weight and cost of the
airborne system [24, 25]. A suitable network protocol is chosen
as required depending upon the range of communication
required. Ground based sensors and controllers send necessary
data to the air segment. Security features of the communication
protocol must be operational for robust and uninterrupted
control of the flying system [26].
500
450
Electrical Energy (MW)
400
350
300
250
200
150
100
50
0
2000
2003
2006
2009
Time
2012
2015
2018
Fig 7. WECS for energy generation and forecast in Pakistan
Figure 8 Map of Pakistan's national grid [32]
As noted from the title of the paper, our target is to address

the energy demands of remote areas of the country far-off from
28
the national grid. Fig. 8 describes the map of Pakistans

national grid which shows large areas of Balochistan not
covered by it. This is because of the low population density in
this province as well as poverty due to which it is not feasible
to extend the grid in these areas. Villages in these areas can be
powered up by renewable energy resources by spending less
money and requiring only maintenance cost in the long run
[33].
TABLE I.
Resource
[7]
[8]
[9]
Energy Demand (M.MTOE)

2008
2022
Oil
11.53
19.18
24.49
Gas Indigenous
11.68
29.88
25.72
8.57
Hydel
4.49
6.86
24.49
Coal
2.10
5.79
18.37
LPG
0.15
0.44
2.45
Nuclear
0.75
3.67
Renewables
14.70
29.94
62.90
122.46
Total
[6]
PAKISTAN ENERGY DEMAND FORECAST [27]
1992
Gas Import
[5]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
Most common uses there require lightening solution, water

pumping for drinkable water supply and also to operate tubewells for agriculture use etc. These applications require
lightweight, low-cost, minimum energy storage solutions.
These demands can be fulfilled by using AWES whereby the
intermittent energy produced during reel out phase could be
used directly.
[18]
[19]
VII. CONCLUSION
[20]
In this paper, an overview of altitude wind technology is

discussed and a preliminary analysis is done for the suitability
to address the needs of renewable energy in developing
countries with budget constraints. A flying wing generator
design is also presented with some initial CFD simulation
results. As the model of FlyPG is now generated, the next step
is to design a prototype system and to demonstrate the energy
output by flying it. Moreover, as the energy needs in the
country are growing, it is high time to invest in such aerospace
technologies to obtain optimal designs resulting in maximum
electrical output with light weight flying systems.
[21]
[22]
[23]
[24]
REFERENCES
[1]
[2]
[3]
[4]
[25]
M. A. Sheikh, "Energy and renewable energy scenario of Pakistan,"

Renewable and Sustainable Energy Reviews, vol. 14, pp. 354-363, 2010.
J. I. Lewis and R. H. Wiser, "Fostering a renewable energy technology
industry: An international comparison of wind industry policy support
mechanisms," Energy Policy, vol. 35, pp. 1844-1857, 2007.
G. Kingsley, "Wind energy production using kites and ground mounted
power generators," ed: Google Patents, 2003.
M. lppolito, "Altitude Wind Energy Generation (Information Dossier),"
KiteGen Research, Italy2007.
[26]
[27]
[28]
29
Proceedings
L. Goldstein, "Theoretical analysis of an airborne wind energy

conversion system with a ground generator and fast motion transfer,"
Energy, vol. 55, pp. 987-995, 2013.
M. I. Blanco, "The economics of wind energy," Renewable and
Sustainable Energy Reviews, vol. 13, pp. 1372-1382, 2009.
B. W. Roberts, "Cost and Security of Electricity Generated by High
Altitude Winds," Altitude Energy2012.
M. Brooks, "To make the most of wind power, go fly a kite," New
Scientist, May, vol. 5, 2008.
R. van der Vlugt, J. Peschel, and R. Schmehl, "Design and Experimental
Characterization of a Pumping Kite Power System," in Airborne Wind
Energy, ed: Springer, 2014, pp. 403-425.
C. Vermillion, B. Glass, and A. Rein, "Lighter-Than-Air Wind Energy
Systems," in Airborne Wind Energy, ed: Springer, 2013, pp. 501-514.
M. Ippolito, "Kite Wind Generator," ed, 2006.
B. Lansdorp and P. Williams, "The laddermill-innovative wind energy
from high altitudes in holland and australia," in Wind Power, 2006.
C. Vermillion, T. Grunnagle, and I. Kolmanovsky, "Modeling and
control design for a prototype lighter-than-air wind energy system," in
American Control Conference (ACC), 2012, 2012, pp. 5813-5818.
G. Mortimer. (2014, 29th October). Multicopter based wind turbines.
Available: http://www.suasnews.com/2010/10/2464/multicopter-basedwind-turbines/
MAKANI. (2013, 20th November). Advanced Wind Energy "ENERGY
KITES". Available: http://www.google.com/makani/
M. Canale, L. Fagiano, and M. Milanese, "High altitude wind energy
generation using controlled power kites," Control Systems Technology,
IEEE Transactions on, vol. 18, pp. 279-293, 2010.
M. Canale, L. Fagiano, M. Ippolito, and M. Milanese, "Control of
tethered airfoils for a new class of wind energy generator," in Decision
and Control, 2006 45th IEEE Conference on, 2006, pp. 4020-4026.
R. Lozano Jr, M. Alamir, J. Dumon, and A. Hably, "Control of a wind
power system based on a tethered wing," Proceedings of second
EGNCA 2012, 2012.
C. Vermillion, B. Glass, and B. Szalai, "Development and Full-Scale
Experimental Validation of a Rapid Prototyping Environment for Plant
and Control Design of Airborne Wind Energy Systems," in ASME 2014
Dynamic Systems and Control Conference, 2014, pp. V002T18A001V002T18A001.
R. Marissen, "Dyneema high strength, high modulus polyethylene
fiber," vol. CIS YA100, Dyneema, Ed., ed: Royal DSM N.V., 2008, p.
4.
L. Fagiano, "Control of tethered airfoils for highaltitude wind energy
generation," Politecnico di Torino, 2009.
M. Ahmed, A. Hably, and S. Bacha, "Grid-connected kite generator
system: Electrical variables control with mppt," in IECON 2011-37th
Annual Conference on IEEE Industrial Electronics Society, 2011, pp.
3152-3157.
P. Williams, B. Lansdorp, and W. Ockesl, "Optimal crosswind towing
and power generation with tethered kites," Journal of guidance, control,
and dynamics, vol. 31, pp. 81-93, 2008.
Z. H. Khan, D. G. Catalot, and J.-M. Thiriet, "Hierarchical wireless
network architecture for distributed applications," in Wireless and
Mobile Communications, 2009. ICWMC'09. Fifth International
Conference on, 2009, pp. 70-75.
Z. H. Khan, J. M. Thiriet, and D. Genon-Catalot, "Wireless network
architecture for diagnosis and monitoring applications," in Consumer
Communications and Networking Conference, 2009. CCNC 2009. 6th
IEEE, 2009, pp. 1-2.
J. Wang, M. Shahidehpour, and Z. Li, "Security-constrained unit
commitment with volatile wind power generation," Power Systems,
IEEE Transactions on, vol. 23, pp. 1319-1327, 2008.
E. A. Council, "Integrated Energy Plan (2009-2022), Report of the
Energy Expert Group " Ministry of Finance2009.
M. A. Khan and U. Ahmad, "Energy demand in Pakistan: a disaggregate
analysis," The Pakistan Development Review, pp. 437-455, 2008.
[29] A. W. Bhutto and S. Karim, "Energy-poverty alleviation in Pakistan

through use of indigenous energy resources," Energy for Sustainable
Development, vol. 11, pp. 58-67, 2007.
[30] M. Hussain, S. Abbas, M. Ansari, A. Zaffar, and B. Jan, "Wind Speed
Analysis of Some Coastal Areas near Karachi," Pakistan Academy of
Sciences, 2012.
[31] I. Ullah, Q.-u.-Z. Chaudhry, and A. J. Chipperfield, "An evaluation of
wind energy potential at Kati Bandar, Pakistan," Renewable and
Sustainable Energy Reviews, vol. 14, pp. 856-861, 2010.
Proceedings
[32] GENI. (2012). Global Energy Network Institute. Available:

http://www.geni.org/globalenergy/library/national_energy_grid/pakistan
/pakistaninationalelectricitygrid.shtml
[33] M. Amer and T. U. Daim, "Selection of renewable energy technologies
for a developing county: a case of Pakistan," Energy for Sustainable
Development, vol. 15, pp. 420-435, 2011.
30
Proceedings
Associate Professor, Dept. of Geo-Informatics,
INTEGRATED USE OF POTENTIAL

RAINWATER HARVESTING SITE
FOR AGRICULTURE USING GEOSPATIAL APPROACH
PMAS Arid Agriculture University Rawalpindi

3
Lecturer Dept. of Geo-Informatics, PMAS Arid
Agriculture University Rawalpindi

4
Lecturer Dept. of Land and Water Conservation,
1,
2,
Niaz Ahmad M.R Khan M.Amin , M.Usman ,
Abida Perveen5
GIS/Hydrology Consultant AAB pvt ltd Islamabad
Professor Dept. of Land and Water Conservation,
agricultural
system
is
critically
dependent
uponsustainable agricultural water management and
its ensured accessibility. Rainwater harvesting has a
greatpotential for storing water for the rising
demands for agricultural purpose. The present
research utilizes Remote Sensing and GIS
applicationsto identify optimal sites for efficient
storage of rainwater for agricultural purposes in the
hilly areas. The essential problem is the optimal
conservation of flows and management of rain water.
We present a valuable methodto improve the
rainwater availability as irrigated water supply to
minimize water deficiency. The main objectives of the
study are: firstly to demarcate the potential

catchments area and then rain-water storage by
proposing reservoir with GIS based estimation of
storage capacity. Aspatial data and non-spatial data is
also generated to plan various agricultural activities
for better utilization of water resources for
agricultural purposes. Spatial database consists of
administrative boundaries, road network, land
use/land cover, Digital Elevation Model (DEM),
settlements points, stream network, catchments area,
proposed water reservoir site, reservoir area and
reservoir capacity. The output of this study can be
used forincreasing the agricultural production due to
efficient utilization of the rainwater.
KeywordsRemote Sensing, GIS, Catchment Area,
Rainwater Harvesting, Spatial Database, DEM,GPS
I.
INTRODUCTION
Water is a limiting factor in Pakistan and land
is not being used according to its potential. It needs
no explanation that there is acute shortage of water
both for drinking and irrigation purposes. Pakistan
is characterized as arid with 80 percent of the land
being arid or semi-arid (Shah, Khan et al. 2012).
Due to steep slopes and impervious nature of
geological formations, a considerable portion of the
runoff goes waste without being utilized for
irrigation and drinking purposes. In addition, the
erosion of top fertile soil has considerably added to
the increased rate of sedimentation in the Indus
River and Arabian Sea. Indus river water quality is
at severe risk from soil erosion. Indus river is one
of the highly polluted rivers (Khan and Ali 2003).
If the present erosion trend continues unabated the
top soil will be completely lost and will result in
vanishing forest with subsequent environmental
repercussions.
Pakistan rain-fed area lies mostly outside Indus
Basin where traditional methods of rainwater
irrigation are practiced. The rainwater besides
traditional methods is regulated through man-made
system like mini dams, small dams, recharging
groundwater and abstracting it through tube-wells,
karezes, wells etc. The Punjab province is blessed
with natural land and water resources(Akhtar
2006). These resources need to be developed for
providing much needed dependable perennial water
supply for irrigated agricultural development of the

area. The development of these resources will also
help towards achieving primary objectives of
socio-economic uplift and poverty alleviation of
the people of the area.
Also the fact due to rapid urbanization,
infiltration of rainwater into the sub-soil has
decreased drastically and reaching groundwater has
diminished. So, the rainwater harvesting is
essential because surface water is inadequate to
meet the ever growing demand and dependence has
to be on groundwater. Rainwater harvesting is a
deliberate collection and storage of rainwater or it
can be said to be an augmentation of groundwater
reservoir by man-made structures by utilizing
rainwater. The objectives and techniques for
rainwater harvesting are highly location specific
and techniques for rainwater harvesting are highly
location specific and an appropriate technology
developed for a particular region cannot be used as
such for other areas for geographic, environmental,
technical and socio-economic reasons(Jasrotia,
Majhi et al. 2009).
RWH can be a measure to increase access
to water for vulnerable sections of the society in
arid and semi-arid parts in countries where
resources are scarce and inaccessible(Munyao,
Mannaerts et al. 2010).
The water harvested can be used for
various purposes ranging from domestic, livestock,
AbstractIn rain fed areas, sustainability of
31
agricultural production, industrial and groundwater

recharge. A successful implementation of RWH
should integrate socio-economic and environmental
issues to ensure sustainability and protect fragile
ecosystems(Vohland and Barry 2009).
RWH may lead to increased food
production through minimizing the risk of crop
failure during droughts and floods. At watershed
level anticipated benefits include recharge to
groundwater systems and improvement of
environment.
The rainwater has to be captured and
made to infiltrate at the place where it falls and
should not be allowed to generate runoff. But, this
cannot be done at all places due to variations in soil
type, geomorphology and other related parameters.
Remote sensing and GIS are very much
helpful in identifying areas which are suitable of
rainwater harvesting. Different soil types have
different capacity for infiltration. Based on their
infiltration rates weights are assigned to them. The
geomorphology of the region is also important in
identifying suitable areas.
There are different hydro-geomorphologic
features with different groundwater potentials,
based on groundwater holding capacity weights are
assigned them too. These weights signify the
suitability of various forms of the same layer. Soil
type and the Hydro geomorphology are two main
criteria involved in identifying suitable areas of
recharge. Hence they are given equal weight-age.
Then suitability indices for each area are
calculated. GIS platform is used for overlaying soil
and hydro geomorphology maps and hence suitable
areas on basis of given weights are identified. The
discussed methodology is suitable for regions with
less rainfall. This technique would help in avoiding
water scarcity during droughts.
Proceedings
between the various factors and existing

technologies.
According
to
Population
Census
Organization and Planning Commission, with the
increasing population of Pakistan estimated to
increase by approximately 50% by the year 2025,
the water needed for meeting food and fiber
requirements will also increase tremendously. To
meet with this increased demand water storages has
to be increased so that more area can be brought
under irrigated agriculture and crop production is
increased. This increase in population will put
additional substantive pressure on the scarce water
resources of Pakistan which are already deficient in
meeting demands of agriculture sector. These
additional resources will relieve the pressure of
water scarcity to some extent for the growing
population. By construction of small water
reservoir the local population near to the reservoirs
will benefit from irrigated agriculture activity.
In this study, we present efficient
approach to mitigate the rainwater while
conserving this floodwater to provide as irrigated
water supply to minimize drought threats. For The
uniqueness of this study is development of a
methodology for estimation of water reservoir
capacity by using remote sensing and GIS tools.
The objectives of paper are i. to demarcate the
potential catchments area ii. rain-water storage by
proposing reservoir with GIS based estimation of
storage capacity. A secondary output, a spatial
catalog is generated for plan to use its spatial data
and non-spatial data for future examine projects.
Spatial database consist of administrative
boundaries, road network, satellite imagery, DEM,
GPS points (GCP), settlements points, stream
network, catchments area, proposed water reservoir
site, reservoir area and reservoir capacity. The
output of this study describe that agricultural water
execution would approved efficiently with the use
of developed methodology.
Most of the decisions on where to locate

the RWH systems are spatial. Different layers of
important factors will be needed to be combined to
determine their suitability. Normally multi-criteria
techniques are used in combining different factors.
In multi criteria evaluation, suitability values of
each factor and their relative importance weights of
different factors need to be established. Suitability
levels refer to the degree to which a certain value in
a given factor influences the location of a rainwater
harvesting technology. For example very steep
slopes will not be suitable areas for bench terraces
compared to gentle slopes. Some RWH techniques
are site-specific and sometimes indigenous
knowledge exists in those sites, therefore, there is a
need for an in depth study at such locations with
the aim of establishing a clear relationship between
the respective factors. For example, Makanya
watershed in Tanzania has relatively high adoption
of various types of RWH technologies. Which
make it an ideal site for studying the relationship
II.
MATERIAL AND METHOD
Location and Accessibility
Study area is located in Punjab region
which is bordered on the west of Khyber
Pakhtunkhwa and part of Balochistan Province; on
the north by Jammu & Kashmir and part of Khyber
Pakhtunkhwa; on the east by Indian part of Punjab
and on south by Sindh Province. It is globally
located between longitudes 69030 to 75010 E and
latitudes 27042 to 34000 N.
Proposed water reservoir is on SirriNullah,
which is tributary of AnkarNullah and joins Soan
River and ultimately falls into Indus River. It is
adjacent to Tamman village in Tehsil Talagang,
District Chakwal. It is about 2.5 miles from
Tamman village. The potential site is approachable
through TalagangTamman road. From Tamman
onwards there is 2 miles metalledand one mile
32
Kacha road. The coordinates of proposed water

reservoir are given below: Northing = 325815.41
Easting = 720840.5
Proceedings
downstream of water reservoir near village

Tamman. Most of the command area is compact
and plain. The area has a gradual slope starting
from the Dam site.
The study area comprises of Potohar
plateau and lies in semi-arid to sub-humid zone of
climatic region with hot summers and cold winters.
The average rainfall at Murree in the north-eastern
part, where the major torrents originate, is over
59.06 inches whereas that of Tamman in the southwestern part of the area is 11.61 inches. About 60%
of the annual rainfall occurs during the Monsoon
season and about 40% in the remaining period.
Generally, the rainfall intensity goes on decreasing
gradually from north-east to south-west. The
average annual rainfall covering the plateau is
estimated at 26.57 inches a year.
Rainfall data of Gujar khan, Jhelum,
Kalabagh and Mianwali has been adopted for the
proposed dam site being nearest rain gauge station
with long term record (1980-2010). Daily rainfall
data of the following stations existing in the
Potohar region is collected from SWHP, WAPDA,
Stations, latitude, longitude; period of record with
aerial distance from study area is labeled in table 2.
Figure 1. Location Map of Study Area
Study area at regional scale consist whole

Potohar Plateau of 22293 km2 covering four
districts Rawalpindi, Chakwal, Attock and Jhelum
of Punjab province of Pakistan. Potential site
selection was proposed through Digital Elevation
Model (DEM) at 3D module of global mapper
software application. Demarcation of catchment
area was done in Arc GIS spatial analyst tool by
using watershed delineation. To find available
rainwater harvesting potential, the region was
divided into sub-catchments using a digital
elevation model through hydrology tools in Arc
GIS. This research was essentially based on field
survey and mapping details. The importance
accumulation of recent research study is that; it is
based on both the use of mapping details and geoinformation tools.
Rainwater Harvesting Mechanism:

Typically rainwater harvesting mechanism
is depicted in rain water accumulate towards low
height grounds. This water normally causes flash
floods but it completely depends upon the localized
terrain. Before allowing it to enter the main stream,
it is preferred to store sufficient amount of water at
local level for agricultural activity This study
assumes that if terrain is mountainous then ideal
solution can be acquired through digital analysis..
This paper clearly suggests that, if we have full
information of the elevations then different kinds of
water reservoir or Dams hence more sustainable
agricultural activity.
Dataset and specification:
The specification of dataset used in the
research work is provided in table1. The procedural
particulars of operations and purpose of use are
discussed under step-wise methodology.
Study area at local scale is proposed on

SirriNullah, which is tributary of Soan River in
district Chakwal and is approachable through
Talagang-Tamman road. From Tamman onwards
there is 2 miles metalled and one mile kacha track.
SirriNullah carries runoff produced by rainfall.
The area is generally characterized by low to
medium relief hills. Highest level in catchment area
is 2100 ftamsl while level of the stream bed at dam
axis is 1133.6 ft-amsl. Dendritic type drainage
pattern is present in catchment area and drainage
pattern is considered to be structurally controlled.
Stream is non-perennial in nature and it is tributary
of Ankarnullah, which is tributary of Soan River.
Soan River joins Indus River upstream of
Kalabagh. Gradient of nullah in the catchment area
is gentle. The catchment area of SirriNullah up to
the proposed water reservoir is 101.47 sq.km. The
proposed command area has about 6000 acre
Table 1: Dataset specification

Data Type
Specification
GPS data
Point data
survey
Landsat TM
30m
Base Map
1:50,000
GeoEye imagery
0.5m resolution
Topography Sheet
Aster
GDEM
Elevation Model
33
1:50,000
Digital
30m
through
ground
Hydrological Map
1:50,000
Climatic data
Point base
Dataset accuracy and quality:

As the key procedure is prepared on
ASTER GDEM based Digital Elevation Model
(DEM), so data authenticity sketch would guide to
ensure the accuracy of output results. The
AsterGDEM enquire elevation data on a nearglobal scale to generate the most complete high
resolution digital topographic database of earth
(Reuter, Nelson et al. 2007)(Rabus, Eineder et al.
2003). Under The base map of the Chakwal and
mountainous area of local level study area provides
the information to identify the target areas of
interest. The map is of large scale i.e. 1:50,000
however for more accuracy assessment was carried
by using GeoEye 0.5m resolution image of
Tamman, Chakwal. The satellite image of Landsat
TM (30 m resolution) was used for land cover
classification. The soil map and hydrological map
were used to digitize the ground water potential
and to identify land productivity classes
respectively at local level study area.
Figure 3. Watersheds of District Chakwal
The catchment area of SirriNullah up to

the proposed dam is 101.47sqkm as shown in
figure 4. The maximum elevation in the catchment

area is 2100 ftamsl while elevation in the river bed
at proposed dam site is 1133.60 ftamsl. The length
of main channel is 17.4 miles and bed slope of Sirri
Figure 2. Major Landcover Map
III.
Proceedings
RESULTS AND DISCUSSION

Figure 4. Catchment area of reservoir
Preparation of GIS Database and Watershed

Delineation
Nullah is 1.05%.There is no perennial
runoff in the nullah. The stream shows a dendritic
stream pattern that is favorable for the surface
runoff produced from rainfall. Map showing land
cover stream meandering at Tamman dam site.
Spatial and non- spatial database is very

significant for GIS based analysis consequently it is
very important to have an accurate geo-database
model
to
utilize
for
GIS
analysis
functionality(Minr, Mentlk et al. 2005). For GIS
data layer we used high resolution satellite
imagery.
For demarcation of Catchment area we
used hydrology tool in spatial analyst extension.
Hydrology tool were used in order as sink-fillcreate flow direction- create- - flow accumulationCreate outlet (pour) points- Delineate Watersheds.
After delineation of watershed raster dataset was
converted to vector for further calculations.
Figure 5.land cover at Tamman dam site

(Source: Google Earth)
34
Proceedings
Rainfall Data
Rainfall data of Kalabagh has been
adopted for the proposed dam site being nearest
rain gauge station with long term record (19872004). Location of climate station is shown in in
table 2 and figure 6. Daily rainfall data of the
following stations existing in the Potohar region is
collected from SWHP, WAPDA. Stations, easting,
northing, period of record with aerial distance from
project area is tabulated below.
Table2 Climatic Stations Location
Eastin
g
Northin
g
EL.(fta
msl)
Aerial
Dist.
from
Project
Area
(miles)
Gujar
Khan
330 15
73018
1476
19622004
68.35
Jhelum
320 56
730 43
755
19612010
92.31
Kalaba
gh
320 57
710 33
702
19872004
31
Mianw
ali
320 35
19621979
43.5
710 31
620
Figure 6. Location of climate station on Google map
MEAN MONTHLY RAINFALL

FOR TAMMAN DAM
RAINFALL (INCHES)
Station
Period
of
Record
Details of Rainfall Analysis

Mean annual rainfall at Jhelum & Gujar Khan
is higher (>32 inches), while Mianwali and
Kalabagh are approximately in between 15-19
inches. Rainfall data of Kalabagh have been
adopted for the proposed dam site because it is very
near (approximately 31 miles) & rainfall in these
areas are approximately same as in the project area.
Daily rainfall data for Mianwali for the period of
1962-1979, Gujar Khan for the period of 19801986 and Kalabagh for the period of 1987-2004
have been collected from Surface Water Hydrology
Project (SWHP), WAPDA. Mean monthly rainfall
at Kalabagh being closest to proposed dam site is
computed for the period 1987-2004. Rainfall
analysis indicated that mean monthly rainfall for
project site varies from 0.26inches to 4.54 inches
(Figure 7) while mean annual rainfall for
corresponding period is 20.32 inches (Figure 8).
5.0
4.5
4.0
3.5
3.0
2.5
2.0
1.5
1.0
0.5
0.0
4.55
0.26
MONTHS
Figure 7. Mean monthly rainfall (inches)
It clearly indicates that proposed dam site lies

between 15-25 inches rainfall contours which are
higher than the rainfall contours at the rainfall
station (19 inch). The rainfall factor therefore
comes to be 22/19 = 1.15. The rainfalls of these
stations are therefore adjusted to derive project
rainfall with factor of 1.15.
35
40
Elevation-Area-Capacity Relationship
MEAN ANNUAL RAINFALL

FOR TAMMAN DAM
Gross Storage Capacity

Gross storage capacity of 8748.80 acre-ft
against El.1212.33 ft Above Mean Sea Level
(amsl) is fixed. Live storage capacity 7025.80 acreft against average annual inflow of 9061 acre-ft
which is above 100% limit.
35
30
RAINFALL (INCHES)
25
Proceedings
Mean Annual
20
Dead Storage Capacity
15
Dead storage is taken as 1723.00 acre-ft

against elevation of 1177.82(amsl). Summary of
these parameters are given in Table 3.
10
5
Table 3. Summary of Reservoir Capacity
Capacity
(AF)
Elevation (ft)
Gross
Capacity
8748.80
1212.33
Dead
Capacity
1723.00
1177.82
Live Capacity
7025.8
YEARS
Figure 8. Annual rainfalls (1962-2005) (inches)
Monthly temperature data for Kalabagh for

period 1986-89 and 1992-2004 have been collected
from SWHP. Monthly maximum and minimum
temperature at Kalabagh are shown in Figure 9.
Data indicates that mean maximum temperature
varies from 23 to 47 centigrade, while mean
minimum temperature varies from 0.3 to 21.5
centigrade.
Month
s
Figure 9.Mean Maximum and Minimum Temperature at
Kalabagh(Degree C)
36
Elevation in ft
TAMMAN DAM SITE

Pond Capacity Curve
REFERENCES
1240
1220
1200
1180
1160
1140
1120
5000
10000
15000
[1]
Minr, J., et al. "Geomorphological information system: idea

and options for practical implementation." Geograficky
Casopis Slovenskej Akademie, Vied57(3), pp. 247, 2005.
[2]
Rabus, B., et al. "The shuttle radar topography missiona

new class of digital elevation models acquired by
spaceborne radar." ISPRS J. Photogrammetry and Remote
Sensing, vol. 57(4), pp. 241-262, 2003.
[3]
Reuter, H. I., et al. "An evaluation of voidfilling

interpolation methods for SRTM data." Intl. J. Geo. Info.
Sci., vol. 21(9), pp. 983-1008, 2007.
Pond Capacity in Acre-Ft
[4]Akhtar, M. R, "Impact of resource conservation technologies for

sustainability of irrigated agriculture in Punjab-Pakistan."
Journal of Agricultural Research (Pakistan), 2006.
Figure 10. Pond Capacity
IV.
Proceedings
CONCLUSIONS
Proposed water reservoir in rain-fed area would have

a significant favourable effect on intended
beneficiaries. In addition to direct benefits from
changes in agricultural landscape there will be
indirect benefits for the relevant households from
appreciation of their land values, creation of
employment
opportunities
/
reduction
in
underemployment and incidence of poverty,
groundwater recharge and improvement in drinking
water supply.
However to ensure realization of anticipated benefits
of this dam, it will be essential to undertake capacity
building of target beneficiaries in terms of financial
resources and knowledge / technology acquisition
regarding high value agricultural commodities /
enterprises.
In fact, the dam site will improve overall water
availability, soil properties, land use and consequent
socio economic status of the inhabitants and thus
enhancing regional development.
Minor possible health considerations will be
countered through inhabitant sensitization and urging
concerned authorities ensuring better healthcare and
regional safeguard measures. The project is thus
highly recommended as Sustainable Development
and could also be replicated to other parts of the
country.
[5]
Jasrotia, A., et al. "Water balance approach for rainwater

harvesting using remote sensing and GIS techniques,
Jammu Himalaya, India." Water resources magt., vol23(14),
pp. 3035-3055, 2009.
[6]
Khan, A. A. and S. B. Ali, "Effects of erosion on Indus

River bio-diversity in Pakistan." Pak. J. Biol. Sci, vol. 6(12),
pp. 1035-1040, 2003.
[7] Minr, J., et al. "Geomorphological information system: idea

and options for practical implementation." Geograficky
Casopis Slovenskej Akademie Vied, 57(3), pp. 247, 2005.
[8]
Munyao, J. N., et al "Use of Satellite Products to Assess

Water Harvesting Potential in Remote Areas of Africa."
Land Degrad. Develop, vol. 22, pp. 359-372, 2010.
[9]
Rabus, B., et al. "The shuttle radar topography missiona

new class of digital elevation models acquired by
spaceborne radar." ISPRS J. Photogrammetry and Remote
Sensing, vol. 57(4), pp.241-262, 2003.
[10]
Reuter, H. I., et al. "An evaluation of voidfilling

interpolation methods for SRTM data." Intl. J. Geo. Info.
Sci,21(9), pp. 983-1008, 2007.
[11]
Shah, H., et al. "Potential For Investment In Indigenous

Technologies: A Case Of Low Cost Soil And Water
Conservation Structures In Rainfed Pothwar, Pakistan."
Pakistan J. Agric. Res, vol. 25(2), 2012.
[12] Vohland, K. and B. Barry, "A review of in situ rainwater

harvesting (RWH) practices modifying landscape functions
in African drylands,"inAgriculture, Ecosystems &
Environment, vol 131(3), pp. 119-127, 2009.
37
Proceedings
Urban change detection of Lahore (Pakistan) using the Thematic Mapper

Images of Landsat since 1992-2010
Institute of Geology, University of the Punjab.
P.O. Box, 54590 Lahore Pakistan.
Email: osman.geomatics@gmail.com
A lot of analytical and statistical models used which was

on based of the urban geometry, spatial relation, social
and economic parameters of a city [2]. In literature there
was three models of urban land use was used (i)
Concentric Zone Model (ii) Sector Model (iii) Multiple
Nuclei.[3] But the Geographic Information System (GIS)
has totally changed the methods to study the urbanization.
Remote sensing with integration of Geographic
information system is now powerful tool to investigate
urban change detection. Traditional method of in field
surveying and mapping required a lot of time span. That
why GIS and remote sensing due to their technical
soundness, are being mostly used for urban change
detection [4]. Integrated study of remote sensing and
geographic information system were used to analyze the
urbanization process of Lahore city using Landsat
Thematic mapper images of 1992-2010. Spatial
information was obtained from the remotely sensed
images of Landsat satellite. While the results shown in the
form of maps are part of geographic information system.
The urban development of Lahore city has been very
rapid during the last decade and this has dramatically
increased the impact on ecosystems. It is therefore
necessary to monitor this expansion. This would be very
helpful for regional development and decision-making
processes. The goal of this study was to detect urban
change in Lahore city using different remote sensing
techniques e,g; Unsupervised classification , urban
change detection, band differencing and red-green
differencing and these techniques reveal that Lahore city
is facing intensely urbanization from 1992 to 2010. The
output is in the form of binary image (called map) which
represents the locations where urban change has occurred.
Abstract: Using the techniques of remote sensing and

geographic information system, this study is an effort to
detect the urbanization in the Lahore (Pakistan) since
1992-2010. Two Landsat TM images from 1992 to 2010
have been used to investigate the urbanization in Lahore.
Using unsupervised classification, we have maked classes
of land cover in each image. Using Erdas imagine 9.2
urban change detection over Lahore has been
investigated. The results of change detection indicate that
urbanization in Shahdera, Kala Shah Kako, Kamoki,
Sheikhopura, Bharia Town and D.H.A regions of Lahore
has increased over 20% since 1992-2010. Band
differencing, Red-Green bands differencing techniques
verify the results of change detection and confirmed these
pockets as intensely urbanized areas since 1992. Maps in
Arc GIS 9.3 show these changes in the images. To
overcome this urbanization this study also guide us about
the Karol,Mohalanwal, Kot Radha Kishan , Kasur etc are
the regions where the new residents can be make.
Keywords: Urban change detection, remote sensing, land
use, Landsat, Urbanization.
I.
Introduction
Urbanization is the conversion of rural areas to urban

settlements in the form of expansion of these settlements.
There are a lot of factors which impact the urbanizations.
The major are industrializations, health, education, and
population increase. Lahore is a historical urban city. It
was located under the old wall city but now it has been
expand tremendously. According to 1998 census eighty
two percent (82%) of the total population of the Lahore
district is urban and eighteen percent (18%) is rural [1].
Spatial variations in urbanization are the variations in the
soil type, habitat, weather pattern and the temporal
variations are the variations which are occurred with time.
Corresponding Author
II.
Contact: +923217579429. Author postal address:

Institute of Geology, University of the Punjab, Lahore
Pakistan
Aim and Objectives
Aim:
To investigate urban change detection of Lahore and
regions around it by using Landsat TM images.
38
Proceedings
Objectives:
(a) To get quality control data of Landsat Thematic
Mapper of year 1990 and 2010.
(b) To investigate and analyze the urban change
detection of Lahore by making the unsupervised
classification.
(c) Validate the areas of urbanization (where
urbanization increase or decrease) by using band
differencing, red green differencing techniques.
(d) Map these areas of urbanization.
III.
Study Area
Lahore (31034 N, 740 22 E) is the capital city of

Punjab province of Pakistan. According to census
1998 it is the second most populated city of Pakistan.
It is located near to the border with India in east,
Sheikhupura on the north-west, Kasur district on the
south. Dated back to a thousand of year, it is an urban
city(see figure 1). This city has extreme climate with
four seasons in a year. Summer season begins in
April and lasted till September. May and June are the
hottest months in the summer where the mean
monthly temperature for maximum and minimum are
45.40 C and 29.30C respectively [4]. Winter season
start from November till March. December is the
coolest month in this season and also coolest month
of the year. The mean monthly maximum and
minimum temperatures for this month are 21.1 0C
and 7.20C respectively [5]. The monsoon period starts
from July and lasted till September. Temperature
falls to 480C in May and June. The average
maximum and minimum temperatures in Lahore are
30.80C and 17.80 respectively. Being a major develop
city, urbanization is going to rise rapidly. It is the
consequences of the coming of people from the
village to city due to more chance of job and better
infrastructure. Lahores urbanization going to
increase since 1973 which is leading to rise in
temperature [5]. Increase in the number of industries
give rise to major source of rise in temperature. Total
area of the Lahore is 1772 km2 [3]. Total population
was 6.319million in 1998 with population density of
3,566 people per km2 [3].
Lahore
Figure 1: Location of Lahore in the Map of Pakistan.

IV.
Methodology
In this study, two Landsat Thematic Mapper images

of October 1992 and October 2010 were acquired
from the USGS website. These images are already
geometrically rectified. The row number for these
images was 149. Thematic Mapper Images mostly
used for the environmental assessment and
monitoring [6]. USGS website provides TM data in
the separate bands. We have performed Layer
stacking on these images. Lahore as the study area
for research, we have cut subsets of both images.
Both subsets images are not of the same size.
October 2010 images subset is more in extending
than October 1992 image (See fig.2 & 3). The band
combination 4, 3, 2 was selected to detect
urbanization [6]. All these processing of data is
performed in Erdas Imagine 9.2 software. To show
results we have made maps in power software of Arc
GIS 9.3
39
Proceedings
performed on software. In which the software search the

pixel of equal brightness value to assigned a class of land
covers. While in the case of Supervised Classifications,
we gives sample class to the software and then software
makes the classes of land covers having same brightness
values of our sample class. We have done with
unsupervised classifications and make four classes in
October 1992 and October 2010 TM images. In October
1992 image these classes are Water, Vegetation, Urban
Area, Unused Land while in October 2010 image, these
are Water, vegetation, Urban Area, Intensely Urbanize
area. Each class was labeled and colored for visual
assessment. (See figure 4 and 5)
Figure 2: Subset of the Landsat TM image showing

Lahore and its neighbors (Oct 1992).
Figure 4: Unsupervised Classification of subset Oct, 1992

(blue is water, green is vegetation, red is urban area).
VI.
Urban change detection is the processing of change

identification in the states of land covers by observing
these changes in the different time periods. With the help
of tool change detection in Erdas imagine 9.2, we have
find out changes occurred the both images (see the figure
6). As we have mentioned before these two subset images
are not of the same size that why borders are in red of
figure 6.
Figure 3: Subset of Landsat TM image of Lahore, October

2010.
V.
Urban Change Detection
Image Classifications
Image classification is generally divided into two broad

categories, Unsupervised Classification and Supervised
Classification. Unsupervised classification is generally
40
Figure 5: Unsupervised classification of TM Oct 2010

image (blue is water, yellow is urban area)
Proceedings
Figure 6: Urban Change Detection of Lahore from

1992 to 2010.
Table 1: Color scheme for the figure 6

Color
Blue
Red
Yellow
Black
Urbanization 1992-2010
5% decrease
> 20% increase
0-20% increase
Back ground in the image
VII.
Band Differencing
With the help of spatial modeler in Erdas Imagine

9.2, we have made a model to find the difference
between two images in band 7 by using nearest
neighbor interpolation technique (See figure 7 and
8).But this model is more suitable for the change
detection visually.
Figure 7: Model in the Erdas Imagine 9.2 for the

band differencing
41
Proceedings
Urbanization
Figure 8: Band differencing of TM of Lahore (19922010).
Figure 9: Red-Green Differencing of Landsat Images

1992-2010.
Light gray areas in the figure 8 indicate the areas where

urbanization has increased [3].This technique utilized the
differences in brightness value of one pixel to
corresponding pixel of another image. Neutral gray tune
in figure 8 black color represent zeros difference while
white has the maximum difference between the images.
VIII.
IX.
Results and Discussions
This study focused on the important of remote sensing

and GIS techniques in the analyzing of urban change
detection of Lahore and its neighbor cities. This study
shows that digital image processing and GIS can be used
for producing urbanized maps of Lahore. As in remote
sensing we have to rely on the number of remote sensing
techniques which leading us to one conclusion. We have
used the TM data with the interval of 18 years to
investigate the urbanizations. Change detection tell us that
during the study period urban sprawl of Lahore has been
shifted towards Shahdera, Kala Shah Kako , Sheikupura
and Kamoki where the urbanization has been increased by
20% . For validation of these results we have use the band
differencing and red-green differencing techniques which
shows that old city of Lahore is going to expand in the
urban areas and this pattern of urbanization is shifted
towards Bharia Town, Johar Town, and Chung. In these
areas people are going to make urbanization over than
20% from 1992-2010. According to [3] Lahore was
58977 hectares in 1972 and which now (2009) has
increased to 99173 hectares. Our results matching with
the results of previous study of Lahores urbanization of >
20% from 1990 to 2009 has occurred. According to 1998
Census, the average family size of the Lahore city was 7.1
Red-Green Differencing
This technique is widely used to find the change in the

areas of two images of different times. To detect the
urban change, open the most recent image (2010) in the
red band and old image (1992) in the green band. In the
Erdas Imagine 9.2, we swipe these two images.
Combination of red and green produce yellow color. But
the areas which have changed show reddish color in
image (see figure 9) are the regions where urbanization
has occurred. In remote sensing we cannot use hard
statements about any results. Through this technique we
also verify the areas where the urbanization has occurred
from 1992-2010.
42
persons, but during 1998 to 2007, the urbanization rate

has increased against population [5].
X.
V. Mesev, 'The Use of Census Data in

Urban
Image
Classification.',
Photogrammetric Engineering and Remote
Sensing, vol. 64, no. 5, pp. 431-438, 1998.
Conclusions
This effort proved its usefulness in the present scenario of

rising trends of population of Lahore city. With the use of
remote sensing and GIS techniques, we have investigated
new pockets around the Lahore where the citizen makes
more residents and plots. Major areas which are facing
urbanization since 1990 are Kala Shah Kako,
Sheikhopura, Meridke, Bheria Town, Kamoki, Chungm
Raiwind, Shahdera and Johar Towns where urbanizations
from 18 years raised 0-20 % even more than 20%. This
study also guides us Karol,Mohalanwal , Kot Radha
Kishan ,Kasur are the regions where the Government can
plan new resident plots. We suggest that future study on
different land use/ land cover patterns of Lahore. Future
study should be undertaken the loss in the area under
decrease in natural vegetation and agriculture. This study
strongly recommend for Government and urban scholars
that such studies should be revised after nominal time
interval with high resolution satellite imagery like Spot,
QuickBird etc.
[5]
[6]
ACKNOWLEDGEMENT
Author would like to say thanks of USGS team who
provide him satellite images for research. Also I would
like to say thanks deeply to Dr. Najam Abbas, Secretary
ICASE 2015,Institute of Space Technology, Islamabad
Pakistan for providing his kind contribution, for
publication of this research work.
XI.
[1]
[2]
[3]
[4]
Proceedings
REFERENCES
District Census Reports for each District

containing General Description of the
District and Broad Analysis of Population
and Housing Data followed by detailed
statistical tables (1998-Census).
Young,X. & Lo, C. P. 'Modelling urban
growth and landscape changes in the Atlanta
metropolitan area', international journal of
geographical science, vol. 17, no. 5, pp.
463,463-488, 2003.
O. Riaz, 'URBAN CHANGE DETECTION
OF LAHORE (PAKISTAN) USING A
TIME SERIES OF SATELLITE IMAGES
SINCE 1972', ASIAN JOURNAL OF
NATURAL & APPLIED SCIENCES., vol. 2,
no. 4, pp. 101,102,103,104, 2013.
Jensen and D. Cowen, 'Remote Sensing of
Urban Suburban Infrastructure and Socioeconomic Attribute', Photogrammetric
Engineering and Remote Sensing, vol. 65,
pp. 611-622, 1999.
43
C. Lo and X. Yang, 'Drivers of land-use/

land-cover changes and dynamics modeling
for the Atlanta. Georgia Metropolitan Area.',
Photogrammetric Engineering and Remote
Sensing,, vol. 68, no. 10, pp. 10731082.,
2002.
ureshi, S. Mahmood, A. S. Almas and R.
Irshad,
'MONITORING
SPATIOTEMPORAL
AND
MICROLEVEL CLIMATIC VARIATIONS IN
LAHORE AND SUBRUBS USING
SATELLITE IMAGERY AND MULTISOURCE DATA', Journal of Faculty of
Engineering & Technology, vol. 19, pp.
53,55, 2012.
H. XU, X. WANG and G. XIAO, 'A
REMOTE
SENSING
AND
GIS
INTEGRATED
STUDY
ON
URBANIZATION WITH ITS IMPACT ON
ARABLE LANDS: FUQING CITY,
FUJIAN PROVINCE, CHINA', LAND
DEGRADATION & DEVELOPMENT, vol.
11, p. 303, 2000.
Proceedings
A Review of Fundamentals of Hyperspectral

Imaging and its Applications
Asad Abbas, Khurram Khurshid
Electrical Engineering Department
Islamabad, Pakistan
Email: asad.abbas@outlook.com, khurram.khurshid@ist.edu.pk
AbstractOver the past two decades, hyperspectral imagery

has emerged as an effective tool for geo-observation using
airborne and satellite based systems. Current sensor technologies
are capable of covering large surfaces of earth with exceptional
spatial, spectral and temporal resolutions. These features allows
the use of hyperspectral imaging in numerous remote sensing
applications requiring estimation of physical parameters of many
complex surfaces and identification of visually similar materials
having fine spectral signatures. Moreover, applications of hyperspectral imaging are growing rapidly, it has shown enormous
potential in various fields including biomedical, archeology, food
quality control, military defense and conservation of artistic and
historic objects. This review describes the differences between
hyperspectral and multispectral data, fundamentals of hyperspectral imaging and focuses on the modern applications of
hyperspectral imagery in precision agriculture, water resource
management and document imaging.
I.
I NTRODUCTION
Human eye is only able to see in a limited part of electromagnetic spectrum and can distinguish between objects based
on their different spectral responses in that narrow spectral
range [1]. However, multispectral imaging sensors have been
developed that are able to acquire an image in infrared and
visible segments of electromagnetic spectrum. Thus allowing
material identification on the basis of their unique spectral signature in a wide spectral range. Multispectral imaging exploits
the property that each material has its own unique spectral
signatures. Spectrum of a single pixel in a multispectral image
provides information about its constituents and surface of the
material.
Multispectral imaging technology is being used for environment and land observational remote sensing from satellite and airborne systems since late 1960s [2]. Multispectral
imaging systems acquire data in a small number of spectral
bands by using parallel sensor arrays. Most of the multispectral
imaging systems uses three to six spectral bands with large
optical band intervals, ranging from visible to near infrared
regions of electromagnetic spectrum for scene observation.
However, such low number of spectral bands are the limiting
factor for discrimination of various materials. The development
of hyperspectral sensing systems over the past two decades
made it possible to acquire several hundred spectral bands
of observational scene in a single acquisition. The increased
spectral resolution of these hyperspectral images allow detail
examination of land surfaces and different materials present
in the observational scene, which was previously not possible
44
with low spectral resolution of multispectral imaging scanners.

Hyperspectral imaging (HSI), also called imaging spectrometer [3], is a technique in which an object is photographed
using several well defined optical bands in broad spectral
range. It was originally implemented on satellite and airborne
platforms for remote sensing applications but during last
two decades, HSI has been applied to numerous applications
including agricultural and water resources control [4], [5],
military defense, art conservation and archeology [6], [7],
medical diagnosis [8], [9], analyses of crime scene details [10],
[11], document imaging [12], forensic medicine [13], food
quality control [14], [15] and mineralogical mapping of earth
surface [16].
This review details the fundamentals of hyperspectral
imaging, discusses the common hyperspectral remote sensing
terminologies and highlights the modern applications of hyperspectral imagery in the field of precision agriculture, water
resource management and document imaging.
II.
F UNDAMENTALS OF H YPERSPECTRAL I MAGING
Unlike normal 2-D images taken by regular camera, hyperspectral images are characterized by their spatial as well
as spectral resolution. The spatial resolution measures the
geometric relationship of the image pixels to each other.
While the spectral resolution determines the variations within
image pixels as a function of wavelength. Table I shows the
spatial and spectral resolution of the current airborne and
space satellite imaging sensors. Due to spectral resolution as
well as spatial resolutions generally hyperspectral images are
referred as hyperspectral data cubes. Figure 1 shows a typical
hyperspectral data cube as a result of hyperspectral imaging.
A. Spatial Resolution
Spatial resolution can be defined as the smallest discernible
detail in an image [17]. Which can be described as the measure
of smallest object in an image, that can be distinguished as a
separate entity in the image. In practical situations clarity of
the image is dictated by it spatial resolution, not the number of
pixels in an image. Spatial characteristics of an image depends
on the design of imaging sensor in terms of its field of view
and its altitude [18]. A finite patch of the ground is captured
by each detector in a remote imaging sensor. Spatial resolution
is inversely proportional to the patch size. The smaller the size
of the patch, higher the details that can be interpreted from the
observed scene.
Proceedings
D. Understanding Spectral Signatures
Fig. 1. Hyperspectral datacube: for any image pixel a complete spectral curve
is observed [12]
TABLE I.
C URRENT S PACEBORNE AND A IRBORNE SYSTEMS

PROVIDING OPTICAL DATA FOR LAND MAPPING
Optical
Spaceborne
Landsat
MODIS
MERIS
ASTER
Hyperion
ALOS
Airborne
AVIRIS
HyMap
ROSIS
DAIS-7915
Subsystem
Spectral
Bands (m)
Spectral
Range (m)
Spatial Resolution
Resolution
Spatial
Coverage
VNIR-TIR
VNIR-TIR
VNIR
VNIR-TIR
VNIR-SWIR
VIS
8
36
15
15
242
1
0.45-12.50
0.40-14.40
0.39-1.040
0.52-11.65
0.40-2.500
0.52-0.77
15-60
250-1000
300
15-90
30
2.5
Global
Global
Global
Global
Regional
Local
VNIR
VNIR-SWIR
VNIR
VNIR-TIR
224
128
115
79
0.38-2.500
0.45-2.480
0.42-0.873
0.45-12
4-20
2-10
2
3-10
Local
Local
Local
Local
B. Spectral Resolution
Spectral resolution can be defined as the number of spectral
bands and range of electromagnetic spectrum measured by
the sensor. An imaging sensor might respond to a large
frequency range but still have a low spectral resolution if it
acquires small number of spectral bands. On the contrary, if
a sensor is sensitive to small frequency range but captures
large number of spectral bands has high spectral resolution,
due to its ability to distinguish between scene elements having
close or similar spectral signatures [19]. Multispectral images
have a low spectral resolution, thus unable to resolve finer
spectral signatures present in the scene. Hyperspectral imaging
(HSI) sensors acquire images in numerous contiguous and
extremely narrow spectral bands in mid infrared, near infrared
and visible segments of electromagnetic spectrum. This type
of advance imaging system shows tremendous potential for
material identification on the basis of their unique spectral
signatures [2]. Spectrum of a single pixel in a hyperspectral
image can give considerably more information about the
surface of the material than a normal image.
C. Temporal Resolution
In hyperspectral remote sensing, the temporal resolution
depends on the orbital characteristics of the imaging sensor. It
is generally defined as the time needed by the sensor platform
to revisit and obtain data from the exact same location [20].
Temporal resolution is said to be high if the revisiting frequency of the sensor platform for the exact same location is
high and is said to be low if revisiting frequency is low. It is
normally defined in days.
45
Materials present on the earths surface emit, transmit,

reflect and absorb electromagnetic energy from the sun in their
own unique way. Hyperspectral sensors allows us to measure
all types of electromagnetic energy within a specified range
as it interacts with materials, thus allowing us to observe the
distinct features and changes on earths surface. Reflectance
is defined as the percentage of amount of energy bouncing
back from materials surface. It is a ratio of reflected energy to
incident energy as a function of wavelength [18]. Reflectance is
100% if all the light energy of specific wavelength striking the
object is reflected back to the imaging sensor, on the other hand
reflectance is 0% if all the incident light of specific wavelength
is absorbed by the object. In most practical cases reflectance
values are in between these two extremes.
In a specified range of electromagnetic spectrum, the
reflectance values of different materials present on the earths
surface such as soil, forest, water, minerals etc can be plotted
and compared. Such plots are labeled as spectral signatures
or spectral response curves [21]. Figure 2 demonstrates a
general model of spectral signatures of different materials
present on the earths surface. Remotely sensed images can
be classified using these spectral signature plots, as each
material present in an observed scene has its own unique
spectral signature. The more the spectral resolution of an
imaging sensor, the more classification information can be
extracted from spectral signatures. Hyperspectral sensors have
high spectral resolution than multispectral sensors and thus
provide the ability to distinguish more subtle differences
in a scene.
Hyperspectral imagery has been utilized by
Fig. 2. A generic scheme of HSI mapping of soil, vegetation and water [22]
geologists for mapping the land and water resources [16]. They
have additionally been utilized to map heavy metals and other
hazardous wastes in historic and active mining areas. Figure 3
shows the spectral response curves of dry bare soil, green
vegetation and clean water.
Figure 3 shows that the reflectance curve for bare soil
has less variations as compared to that of vegetation. This is
because of the fact that the factors that affect soil reflectance
vary in a narrow range of electromagnetic spectrum. These
factors include soil texture, presence of minerals such as iron,
surface roughness and moisture content in soil [21]. Spectral
Fig. 3.
Spectral response curves of soil, vegetation and water [18]
signatures of green vegetation have valleys in the visible

portion of spectrum that indicates the pigmentation of the
plant. Chlorophyll is the primary photosynthetic pigment in
green vegetation [18], it is absorbs strongly in red (670 nm)
and blue (450 nm) regions called the chlorophyll absorption
spectral bands. When a plant is under stress so that the
chlorophyll growth is reduced, in such cases the amount of
reflectance in red (670 nm) regions increases [18]. The spectral
response of water has a distinctive characteristics of absorption
of light in near infrared and beyond. Common factors affecting
spectral response of water are the suspended sediments and
increases in chlorophyll levels. In each cases spectral response
will be shifted accordingly showing the presence of suspended
sediments or algae in water [21].
III.
A PPLICATIONS OF H YPERSPECTRAL I MAGING
A. Precision Agriculture
Many studies have indicated that the worlds crop
production needs to be doubled by the end 2050 due to
the rapid increase in worlds population [23]. However,
various studies have shown that the crop yields are not longer
increasing at a rate to fulfill the growing population needs
[24], [25]. Recent studies have indicated that increasing
crop yields, rather than using more land for cultivation,
is the most effective path for ensuring food security [26],
[27]. Global poverty and undernourishment can directly be
reduced by increasing crop production, moreover most of
the poor and undernourished population consists of farmers
themselves [28].
Traditionally crop monitoring for disease, water stress,
nutrients and insect attack was carried out by manual visual
inspection from the ground. These methods were limited by
the fact that the visual symptoms often appear at later stages
of disease, thus making it difficult to restore plant health.
Advancement in airborne and ground based hyperspectral
imaging methods has made possible the evaluation of crop
stresses, analyzing soil and vegetation characteristics in a
cost effective manner, thus replacing the traditional scouting
methods.
Drought stress is an important factor affecting crop
yields. Chances of a successful crop can be highly increased
by timely detection of water related stresses. High water
level stresses are noticeable in variations in photosynthetic
pigments. These changes leads to yellowish tint in crops, due
to the increase reflectance of red wavelength. Unlike human
46
Proceedings
eye, hyperspectral imaging sensors can detect these changes

at earlier stages. Colombo et al. [29] indicated that changes
in leaf equivalent water thickness (EWT) was responsible
for changes in leaf reflectance in the infrared and visible
spectrum. They stated that hyperspectral regression indices
calculated from hyperspectral imaging were powerful tools
for estimation of water content at leaf as well as at landscape
level. Rascher et al. [30] used a portable hyperspectral
imaging system and photochemical reflectance index to
estimate water stress in leaves of tropical tress and observed
the temporal effects of dehydrations on tree leaves. Rossini
et al. [31] found that hyperspectral imaging is useful in
detecting drought stress at farm level with corn. They showed
that irrigation deficits can be accurately mapped well before
drought stress affected the canopy structure.
Deficiencies in nutrients and soil contamination causes
various symptoms that can be assessed by hyperspectral
imaging. Schuerger et al. [32] used hyperspectral imaging to
observe zinc deficiency and toxicity to determine chlorophyll
levels relating to stress symptoms. They indicated that
traditional direct sampling methods are much more costly
than hyperspectral imaging. Dunagen et al. [33] analyzed
mercury levels in mustard plants and found that spectral
signatures were notably related to the contaminant levels.
Osborne et al. [34] showed that biomass, yield under
stress, nitrogen and phosphorous concentrations can be
estimated by using hyperspectral imaging. Mahlein et al. [35]
studied different development stages of diseased suger beet
leaves using hyperspectral imaging. Figure 4 shows spectral
signatures of healthy as well as diseased sugar beet leaves.
This study also showed that hyperspectral imaging has a great
potential for analyzing plant diseases.
Analysis of soil characteristics can play a vital role in
Fig. 4. Spectral signatures comparison of healthy and diseased suger beet

leaves [35]
increasing crop yields. Ben-Dor et al. [36] al successfully

mapped vital characteristics of soil in a field scale experiment
including moisture, soil organic matter and soil sanity. Rossel
and Bratney [37] estimated the organic carbon content in soil
with accuracy.
B. Water Resource Management
Water is one of the most important resource available
on the earth and is vital for survival of humanity. For this
reason, managing the water resource efficiently, analyzing

and monitoring the quality of water has attracted a lot of
attention from the researchers [38], [39], [40], [41], [42], [43].
Hyperspectral remote sensing technology has found enormous
applications in water resource management. Accurate estimates
of water resource parameters are possible by analyzing spatial,
spectral and temporal variations in water bodies.
The efficacy of flood detection and monitoring system is
limited by their incapacity to get important information about
water conditions from airborne and ground observatories in a
timely manner. Recent improvements in remote sensing technology has improved the early flood warning system and vastly
reduced the time of detection and reaction to flood events
to a few hours [44]. US geological survey and NASA are
incorporating space borne observations of rainfall resources,
rivers and land topography into early warning systems with
potential global applications [45]. Glaber and Reinartz [46]
studies the optimal procedure for detection of flooded areas
with remote sensing data. They investigated the erosive impact
of floods, moisture content in flood plain areas, accumulation of sediments. Roux and Dartus [47] explored the flood
hydrographs and estimated the river discharge from remotely
sensed data. They optimized their model to minimize the error
between system response and their proposed model to estimate
the river discharge.
Hyperspectral remote sensing provides efficient and reliable information about water quality parameters which contain
biochemical, hydro-physical and biological attributes [48] [49]
[50]. Hyperspectral imaging enable us to measure chlorophyll,
turbidity and chemical oxygen demand and phosphorous in
water resources. Chlorophyll content in water is extensively
studied by hyperspectral remote sensing, which gives and
estimate of algal level and hence water quality. Studies have
been carried out for evaluation of ammonia changes for
wetland [51], classifying different parameters of lakes [52],
estuaries [53] and analyzing algal blooms [54].
Wetland mapping has played a significant role in order
to enhance the quality of our ecosystem [55]. Hyperspectral
imagery has helped in detailed understanding of vegetation
characteristics of ecosystem. Extensive research studies have
been carried out using remote sensing to explore the significance of acquiring timely data for monitoring and mapping
aquatic vegetation [56], which is said to be an important aspect
in ecosystem reconstruction and restoration.
Proceedings
readability [58], ink aging and forensic document analysis [59].

Hyperspectral document imaging works on the principle that
each ink present in the document has its own unique spectral
signature. Many mathematical tools that are being used in
hyperspectral remote sensing can be used on hyperspectral
document images for classification, improving legibility of
extremely deteriorated text, ink aging and fraud detection.
Moreover hyperspectral imaging is non-destructive, automated
and environment insensitive tool for document examination.
In hyperspectral document imaging, ink mismatch detection analysis provides important information to forensic
document examiners to determine the authenticity of legal
documents. Forgery, backdating and fraud can be detected
using ink analysis of documents. Recently, we have worked on
a non-destructive automated hyperspectral document imaging
analysis system, which was able successfully discriminates
between different visually similar ink mixtures (Dataset D1 to
D7) taken from hyperspectral imaging database [60]. Figure 5
shows the final segmentation results of our proposed approach.
C. Document Imaging
Traditionally, forensic document experts and paleographers
used chemical solution based methods to study the extrinsic
and intrinsics components of the important historic documents [57]. This is due to the fact that the inks used on
documents throughout the history were composed of diverse
substances having distinct chemical and physical properties.
Each of these substances have their own unique way for
reacting with different substrates depending upon the reaction
environment. These chemical solution based methods helped
in document analysis. But unfortunately these techniques were
time consuming, sensitive to temperature changes and destructive in nature i.e. harms to the important documents were
irreversible.
To overcome such limitations, hyperspectral imaging has
emerged as an effective non-destructive tool for improving
47
Fig. 5. Comparisons of Ground Truth Images (Dataset D1 to D7) with final

segmentation results
IV.
D ISCUSSION AND C ONCLUSION
Among remote sensing technologies, the role of hyperspectral imagery in the geo-observation, identification and detection
of materials and estimation of physical parameters cannot be
stated enough. Due to this very reason there are increasing
number of airborne and spaceborne hyperspectral platforms
based applications being researched. Recent advancement in
sensor technologies has encouraged researchers to use hyperspectral imagery in many modern applications. Many mathematical tools and algorithms are being researched such as data
fusion, hyperspectral unmixing, hyperspectral classification,
anomaly detection and fast computing for efficient utilization
of hyperspectral data. These mathematical tools can be used
on hyperspectral data across many different applications.
In general, this review focuses on the vast extent to which
hyperspectral imaging has been used to increase the crop
yields, managing water resources and its advanced application
in historic and modern document imaging. Promising results
have been found in those areas and also future research is
being carried out for further improvement as well.
[16]
R EFERENCES
[22]
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
E. H. Land and J. McCann, Lightness and retinex theory, JOSA,

vol. 61, no. 1, pp. 111, 1971.
D. Landgrebe, Information extraction principles and methods for
multispectral and hyperspectral image data, Information processing for
remote sensing, vol. 82, pp. 338, 1999.
J. Solomon and B. Rock, Imaging spectrometry for earth remote
sensing, Science, vol. 228, no. 4704, pp. 11471152, 1985.
M. Govender, K. Chetty, and H. Bulcock, A review of hyperspectral
remote sensing and its application in vegetation and water resource
studies, Water Sa, vol. 33, no. 2, 2007.
E. Adam, O. Mutanga, and D. Rugege, Multispectral and hyperspectral
remote sensing for identification and mapping of wetland vegetation: a
review, Wetlands Ecology and Management, vol. 18, no. 3, pp. 281
296, 2010.
C. Fischer and I. Kakoulli, Multispectral and hyperspectral imaging
technologies in conservation: current research and potential applications, Studies in Conservation, vol. 51, no. Supplement-1, pp. 316,
2006.
H. Liang, Advances in multispectral and hyperspectral imaging for
archaeology and art conservation, Applied Physics A, vol. 106, no. 2,
pp. 309323, 2012.
O. Carrasco, R. B. Gomez, A. Chainani, and W. E. Roper, Hyperspectral imaging applied to medical diagnoses and food safety, in AeroSense
2003. International Society for Optics and Photonics, 2003, pp. 215
221.
M. A. Afromowitz, J. B. Callis, D. M. Heimbach, L. A. DeSoto, and
M. K. Norton, Multispectral imaging of burn wounds: a new clinical
instrument for evaluating burn depth, Biomedical Engineering, IEEE
Transactions on, vol. 35, no. 10, pp. 842850, 1988.
J. Kuula, I. Polonen, H.-H. Puupponen, T. Selander, T. Reinikainen,
T. Kalenius, and H. Saari, Using vis/nir and ir spectral cameras for
detecting and separating crime scene details, in SPIE Defense, Security,
and Sensing. International Society for Optics and Photonics, 2012, pp.
83 590P83 590P.
R. L. Schuler, P. E. Kish, and C. A. Plese, Preliminary observations
on the ability of hyperspectral imaging to provide detection and visualization of bloodstain patterns on black fabrics, Journal of forensic
sciences, vol. 57, no. 6, pp. 15621569, 2012.
R. Padoan, T. A. Steemers, M. Klein, B. Aalderink, and G. De Bruin,
Quantitative hyperspectral imaging of historical documents: technique
and applications, ART Proceedings, 2008.
G. Edelman, E. Gaston, T. Van Leeuwen, P. Cullen, and M. Aalders,
Hyperspectral imaging for non-contact analysis of forensic traces,
Forensic science international, vol. 223, no. 1, pp. 2839, 2012.
A. Gowen, C. ODonnell, P. Cullen, G. Downey, and J. Frias, Hyperspectral imagingan emerging process analytical tool for food quality
and safety control, Trends in Food Science & Technology, vol. 18,
no. 12, pp. 590598, 2007.
Y.-Z. Feng and D.-W. Sun, Application of hyperspectral imaging in
food safety inspection and control: a review, Critical reviews in food
science and nutrition, vol. 52, no. 11, pp. 10391058, 2012.
48
[17]
[18]
[19]
[20]
[21]
[23]
[24]
[25]
[26]
[27]
[28]
[29]
[30]
[31]
[32]
[33]
[34]
[35]
Proceedings
R. N. Clark and G. A. Swayze, Mapping minerals, amorphous materials, environmental materials, vegetation, water, ice, and snow, and
other materials: The usgs tricorder algorithm, in Summaries of the Fifth
Annual JPL Airborne Earth Science Workshop, vol. 1. JPL Publication,
1995, pp. 3940.
R. C. Gonzalez, Digital image processing. Pearson Education India,
2009.
R. Smith, Introduction to remote sensing of the environment, 2001b.
www. microimages. com, 2001.
J. R. Jensen, Introductory digital image processing: A remote sensing
perspective. Pearson College Division, 2005.
J. Thau, Temporal resolution, in Encyclopedia of GIS. Springer US,
2008, pp. 11501151. [Online]. Available: http://dx.doi.org/10.1007/
978-0-387-35973-1 1376
T. Lillesand, R. W. Kiefer, and J. Chipman, Remote sensing and image
interpretation. John Wiley & Sons, 2014.
G. A. Shaw and H.-h. K. Burke, Spectral imaging for remote sensing,
Lincoln Laboratory Journal, vol. 14, no. 1, pp. 328, 2003.
D. Tilman, C. Balzer, J. Hill, and B. L. Befort, Global food demand
and the sustainable intensification of agriculture, Proceedings of the
National Academy of Sciences, vol. 108, no. 50, pp. 20 26020 264,
2011.
K. G. Cassman, Ecological intensification of cereal production systems: yield potential, soil quality, and precision agriculture, Proceedings of the National Academy of Sciences, vol. 96, no. 11, pp. 5952
5959, 1999.
D. K. Ray, N. Ramankutty, N. D. Mueller, P. C. West, and J. A.
Foley, Recent patterns of crop yield growth and stagnation, Nature
communications, vol. 3, p. 1293, 2012.
H. C. J. Godfray, J. R. Beddington, I. R. Crute, L. Haddad, D. Lawrence,
J. F. Muir, J. Pretty, S. Robinson, S. M. Thomas, and C. Toulmin, Food
security: the challenge of feeding 9 billion people, science, vol. 327,
no. 5967, pp. 812818, 2010.
J. A. Foley, N. Ramankutty, K. A. Brauman, E. S. Cassidy, J. S. Gerber,
M. Johnston, N. D. Mueller, C. OConnell, D. K. Ray, P. C. West et al.,
Solutions for a cultivated planet, Nature, vol. 478, no. 7369, pp. 337
342, 2011.
P. L. Pingali, Green revolution: Impacts, limits, and the path ahead,
Proceedings of the National Academy of Sciences, vol. 109, no. 31, pp.
12 30212 308, 2012.
R. Colombo, M. Meroni, A. Marchesi, L. Busetto, M. Rossini, C. Giardino, and C. Panigada, Estimation of leaf and canopy water content
in poplar plantations by means of hyperspectral indices and inverse
modeling, Remote Sensing of Environment, vol. 112, no. 4, pp. 1820
1834, 2008.
U. Rascher, C. J. Nichol, C. Small, and L. Hendricks, Monitoring
spatio-temporal dynamics of photosynthesis with a portable hyperspectral imaging system, Photogrammetric Engineering & Remote Sensing,
vol. 73, no. 1, pp. 4556, 2007.
M. Rossini, F. Fava, S. Cogliati, M. Meroni, A. Marchesi, C. Panigada,
C. Giardino, L. Busetto, M. Migliavacca, S. Amaducci et al., Assessing
canopy pri from airborne imagery to map water stress in maize, ISPRS
Journal of Photogrammetry and Remote Sensing, vol. 86, pp. 168177,
2013.
A. C. Schuerger, G. A. Capelle, J. A. Di Benedetto, C. Mao, C. N.
Thai, M. D. Evans, J. T. Richards, T. A. Blank, and E. C. Stryjewski,
Comparison of two hyperspectral imaging and two laser-induced
fluorescence instruments for the detection of zinc stress and chlorophyll
concentration in bahia grass (paspalum notatum flugge.), Remote
sensing of environment, vol. 84, no. 4, pp. 572588, 2003.
S. C. Dunagan, M. S. Gilmore, and J. C. Varekamp, Effects of mercury
on visible/near-infrared reflectance spectra of mustard spinach plants
(brassica rapa p.), Environmental Pollution, vol. 148, no. 1, pp. 301
311, 2007.
S. Osborne, J. S. Schepers, D. Francis, and M. R. Schlemmer, Detection of phosphorus and nitrogen deficiencies in corn using spectral
radiance measurements, Agronomy journal, vol. 94, no. 6, pp. 1215
1221, 2002.
A.-K. Mahlein, U. Steiner, C. Hillnhutter, H.-W. Dehne, and E.-C.
Oerke, Hyperspectral imaging for small-scale analysis of symptoms
caused by different sugar beet diseases, Plant methods, vol. 8, no. 1,

p. 3, 2012.
[36]
E. Ben-Dor, K. Patkin, A. Banin, and A. Karnieli, Mapping of several

soil properties using dais-7915 hyperspectral scanner data-a case study
over clayey soils in israel, International Journal of Remote Sensing,
vol. 23, no. 6, pp. 10431062, 2002.
[37]
C. Gomez, R. A. V. Rossel, and A. B. McBratney, Soil organic

carbon prediction by hyperspectral remote sensing and field vis-nir
spectroscopy: An australian case study, Geoderma, vol. 146, no. 3,
pp. 403411, 2008.
[38]
F. Zagolski, V. Pinel, J. Romier, D. Alcayde, J. Fontanari, J. GastelluEtchegorry, G. Giordano, G. Marty, E. Mougin, and R. Joffre, Forest
canopy chemistry with high spectral resolution remote sensing, International Journal of Remote Sensing, vol. 17, no. 6, pp. 11071128,
1996.
[39]
G. P. Asner, Biophysical and biochemical sources of variability in

canopy reflectance, Remote sensing of Environment, vol. 64, no. 3, pp.
234253, 1998.
[40]
K. McGwire, T. Minor, and L. Fenstermaker, Hyperspectral mixture

modeling for quantifying sparse vegetation cover in arid environments,
Remote Sensing of Environment, vol. 72, no. 3, pp. 360374, 2000.
[41]
C. Stone, L. Chisholm, and N. Coops, Spectral reflectance characteristics of eucalypt foliage damaged by insects, Australian Journal of
Botany, vol. 49, no. 6, pp. 687698, 2001.
[42]
N. Coops, S. Dury, M.-L. Smith, M. Martin, and S. Ollinger, Comparison of green leaf eucalypt spectra using spectral decomposition,
Australian journal of botany, vol. 50, no. 5, pp. 567576, 2002.
[43]
E. Underwood, S. Ustin, and D. DiPietro, Mapping nonnative plants

using hyperspectral imagery, Remote Sensing of Environment, vol. 86,
no. 2, pp. 150161, 2003.
[44]
F. Ip, J. Dohm, V. Baker, T. Doggett, A. Davies, R. Castano, S. Chien,

B. Cichy, R. Greeley, R. Sherwood et al., Flood detection and
monitoring with the autonomous sciencecraft experiment onboard eo-1,
Remote Sensing of Environment, vol. 101, no. 4, pp. 463481, 2006.
[45]
G. Brakenridge, E. Anderson, and S. Nghiem, Satellite microwave

detection and measurement of river floods, in AGU Spring Meeting
Abstracts, vol. 1, 2006, p. 05.
[46]
C. Glaer and P. Reinartz, Multitemporal and multispectral remote

sensing approach for flood detection in the elbe-mulde region 2002,
Acta hydrochimica et hydrobiologica, vol. 33, no. 5, pp. 395403, 2005.
[47]
H. Roux and D. Dartus, Use of parameter optimization to estimate a

flood wave: Potential applications to remote sensing of rivers, Journal
of Hydrology, vol. 328, no. 1, pp. 258266, 2006.
[48]
M. Chawira, T. Dube, and W. Gumindoga, Remote sensing based

water quality monitoring in chivero and manyame lakes of zimbabwe,
Physics and Chemistry of the Earth, Parts A/B/C, vol. 66, pp. 3844,
2013.
[49]
T. Kutser, B. Paavel, C. Verpoorter, T. Kauer, and E. Vahtmae, Remote

sensing of water quality in optically complex lakes, in Proceedings of
the XXII Congress of the International Society for Photogrammetry and
Remote Sensing, Melbourne, Australia, 2012.
[50]
S. Koponen, Remote sensing of water quality: the development and

use of water processors available in beam, Finland, Helsinki: Helsinki
University of Technology TKK, Department of Radio Science and
Engineering, pp. 158, 2006.
[51]
D. R. Tilley, M. Ahmed, J. H. Son, and H. Badrinarayanan, Hyperspectral reflectance of emergent macrophytes as an indicator of water
column ammonia in an oligohaline, subtropical marsh, Ecological
Engineering, vol. 21, no. 2, pp. 153163, 2003.
[52]
S. Thiemann and H. Kaufmann, Lake water quality monitoring using

hyperspectral airborne dataa semiempirical multisensor and multitemporal approach for the mecklenburg lake district, germany, Remote
sensing of Environment, vol. 81, no. 2, pp. 228237, 2002.
[53]
J.-M. Froidefond, L. Gardel, D. Guiral, M. Parra, and J.-F. Ternon,

Spectral remote sensing reflectances of coastal waters in french guiana
under the amazon influence, Remote Sensing of Environment, vol. 80,
no. 2, pp. 225232, 2002.
[54]
R. P. Stumpf, Applications of satellite ocean color sensors for monitoring and predicting harmful algal blooms, Human and Ecological Risk
49
[55]
[56]
[57]
[58]
[59]
[60]
Proceedings
Assessment: an International Journal, vol. 7, no. 5, pp. 13631368,

2001.
K. Schmidt and A. Skidmore, Spectral discrimination of vegetation
types in a coastal wetland, Remote sensing of Environment, vol. 85,
no. 1, pp. 92108, 2003.
L. Yuan and L. Zhang, Identification of the spectral characteristics of
submerged plant vallisneria spiralis, Acta Ecologica Sinica, vol. 26,
no. 4, pp. 10051010, 2006.
V. Aginsky, Forensic examinaton of slightly soluble ink pigments
using thin-layer chromatography, Journal of Forensic Sciences, vol. 38,
pp. 11311131, 1993.
S. J. Kim, F. Deng, and M. S. Brown, Visual enhancement of old
documents with hyperspectral imaging, Pattern Recognition, vol. 44,
no. 7, pp. 14611469, 2011.
G. Edelman, E. Gaston, T. Van Leeuwen, P. Cullen, and M. Aalders,
Hyperspectral imaging for non-contact analysis of forensic traces,
Forensic science international, vol. 223, no. 1, pp. 2839, 2012.
A. Abbas, K. Khurshid, and F. Shafait, Ink Mismatch Detection using
Hyperspectral Document Image Analysis, Masters thesis, Institute of
Space Technology, Islamabad, Pakistan, 2015.
Proceedings
A Numerical Study on the Impact Resistance of

Braided Composites
X. Lei
Dept. of Mechanical
Eng., Hanyang
University, South
Korea
K. Hayat
Dept. of Mechanical Eng.,
University of Lahore, Lahore,
Pakistan
Email: khazar.hayat@me.uol.edu.pk
M. T. Hussain
Satellite Research
and Development
Center, Lahore,
Pakistan
H. T. Ali
Dept. of Aerospace Eng.,
Queen's Building,
University of Bristol,
BS8 1TR, UK
R. S. Choudhry
Dept. of Mechanical Eng.,
National University of
Science and Technology,
Islamabad, Pakistan
transformation of stresses and effective material properties

between the macro level and micro level analyses.
Abstract Braided fabric reinforced composites are

attractive alternatives to conventional prepreg laminates due to
higher impact resistant performance. For the development of
aerospace structures made of braided composites, the numerical
simulations can play a key role in reducing time and cost.
However, the numerical tools to evaluate the impact response and
to investigate the damage mechanism of these materials, are not
fully developed yet. This work focuses on the numerical
modelling and analysis of the impact induced response of braided
composites under different impact velocities, using micromechanics of failure (MMF) together with progressive damage
models for fibre and matrix constituents. For implementation of
MMF, the meso- and micro-scale models of unit cell have been
utilized. In the meso-scale unit cell of the braided composites, the
tows effective material properties were updated using a micro
unit cell for which the degradation of constituent material
properties was incorporated. The impact induced damage
progression for the tows and pure matrix was then evaluated.
The impact analyses were performed for bi-axial braided
composites at 45 degrees and 25 degrees braiding angles to
demonstrate the feasibility of the MMF-based approach.
Moreover, the impact response of the braided composites with
both thermosetting and thermos-plastic resin system was also
investigated.
II. MICROMECHANICS-BASED PROGRESSIVE DAMAGE MODEL

A. Geometric modeling of braided composites
The development of an idealized model of the braided
composites is necessary to predict their material behaviors.
The fiber bundle arrangements are modelled using the
geometrical parameters (i.e. braiding angle, width and
thickness of axial tow, width and thickness of bias tow, and
gap between bias tows) illustrated in Fig. 2. Moreover, in the
modeling, the undulation of tows plays a pivotal role in
determining the behavior of the braided composites. Further
details on modeling procedure and techniques can be found in
[2].
Keywords Braided composites, Impact resistance, Micromechanics of failure, Progressive damage model
I. INTRODUCTION
Recently, a rapid growth of the applications of textile
braided fabric reinforced composites have been observed in
the marine, aerospace and automobile industries etc. due to its
net-shape fabrication and resistance to delamination damage
and impact load in particular [1]. However, the behavior of the
braided composites is not fully understood yet, and there is a
need of establishing a methodology that can predict the
material failure behaviors under different types of loadings,
primarily the impact loading [2].
Fig. 1. Analysis of braided composites.
The purpose of this study is to evaluate the behavior of

braided composites subjected to impact load with
thermoplastic and thermosetting resin systems, based on
various braiding angles and using micromechanics of failure
(MMF) theory [3]. The analysis procedure involves three
scopes: macro, meso and micro, as shown in Fig. 1. A braided
composite can be assumed to be limited to mesoscale. It
consists of tows and pure matrix encapsulating the tows. Thus,
the use of mesoscale unit cell model serves as a bridge for
Fig. 2. Modeling of braided composites.
B. Meso and micro analyses

The meso- and micro-level analyses are bridged to
characterize the material behaviour of the tows of the braided
50
composites. It should be noted that a tow can be approximated

as a curved continuous unidirectional (UD) lamina, and a
micro unit cell, shown in Fig. 3c, can be used to relate the
constituent behaviour to the tow behaviour. Thus, tow stresses
can be transformed into each constituent stresses, and resulting
constituent damages can be used to determine the tow damage.
For the constituent damage analysis, the degradation of the
constituent stiffness properties is carried out, and the tow
effective material properties are subsequently updated through
the micro unit cell.
Stassis equivalent stress

strength :
Proceedings
, reaches the matrix tensile
(5)
where the symbol represents the ratio of the compressive

strength to the tensile strength of the matrix constituent.
Further, the equivalent stress can be transformed into the
equivalent strain of the matrix constituent, described in Eq. 6,
using the stress-strain relations.
(6)
where ,
, and represents the first strain invariant, the
von Mises equivalent strain and the Poissons ratio of the
matrix constituent.
Fig. 3. Estimation of effective material properties of tows of braided
composites.
It should be noted that the equivalent strain, like the

equivalent stress, is a scalar quantity that is computed from the
strain components. Fig. 4a represents a multi-linear stressstrain model defined using the equivalent stress and the
equivalent strain, to estimate the matrix constituent damage.
The matrix constituent behaves in a linear manner, before the
occurring of any damage. Once damage occurs, the matrix
constituent demonstrates a nonlinear behavior (i.e. hardening
followed by the softening depending on the damage status).
The stiffness of the matrix constituent is degraded by a factor
as following:
The transformation between the tows stresses at mesolevel, denoted by

, and the fiber and matrix stresses at
micro-level, denoted by , and
, is carried using the stress
amplification factors (SAFs), denoted by and matrices, as
following [3, 4]:
(1)
Eq. (1) can be written in detailed form, as following [4]:
(7)
subjected to the condition:

. The symbols
and
denotes the matrix yield stress and the matrix
(2)
and
yield strain at the beginning, and the symbols
denotes the matrix yield stress and matrix yield strain at the
denotes the current equivalent
end of ith step. The symbol
strain of the matrix at the ith step, and the symbol
denotes
the matrix intact stiffness.
In a similar manner, the meso strain and temperature

increment can also be transformed to micro strains, as shown
in Eq. 3, using the strain amplification factors which can be
derived from SAFs using the constituent compliance matrix
and the unit cell compliance matrix .
For the mesoscale model, the damage in the pure matrix

was modeled using Eq. 8 only, and additional MMF-based
damage modeling using a micro unit cell was used for the
matrix of tows [6]. The periodic boundary conditions were
used for the meso and micro unit cells [2, 7]. The micro unit
cell model, based on the MMF theory and linear material
model, delivered multi-axial micro stress state for the fibre
and the matrix constituents (i.e. assuming perfect bonding for
interface). To avoid computation instability, the maximum
local damage value for fiber and matrix constituent was set to
0.9 [8].
(3)
C. Damage model of micro-constituents

For each constituent, a separate failure criterion is used
due to their distinct mechanical behaviors. Considering the
isotropic nature of the matrix constituent, the von Mises
failure criterion [3, 5] is used, which follows:
(4)
The mesoscale stresses were applied on the micro unit

cell. After performing the finite element analysis, the micro
stresses/strains in the matrix were computed, as shown in Fig.
5. Afterwards, the damage analysis for the matrix was
performed. For the matrix of tow, the localization of damage
where ,
,
, and
represents the first stress invariant,
von Misses equivalent stress, tensile strength and compressive
strengths of matrix constituent, respectively. The failure
criterion described by Eq. 4 matches the condition that the
51
Proceedings
was avoided by using a volume-based damage

homogenization technique, represented by Eq. 8 shown below:
(10)
(8)
denotes the damage factor of jth

where the symbol
element of the fiber constituent. And, the constitutive relation
for fiber constituent was:
(11)
where the symbol denotes a positive weighing factor,

and the symbol denotes the matrix overall volume in the
micro unit cell.
where
denotes the stiffness of overall fiber zone in the
micro unit cell.
D. Numerical implementation
Fig. 6 illustrates the flow chart of the algorithm developed
for combining the MMF methodology and progressive damage
model. In the beginning, the total global strain, denoted
by
, is computed at the global time , and augmenting the
global strain increment, denoted by
, to global strain at
previous time-step n-1. For the tows, the macro stresses of
each tow element, denoted by
, are estimated using the
.
previous effective stiffness properties, denoted by
Afterwards, the micro stresses for each element of the both
matrix and fiber zones in the micro unit cell, denoted by
, are computed from the meso stress, denoted by
,
and
by using SAFs. The constituent failure criteria are then applied
to both matrix and fiber, and damage factor is estimated for
each jth element in the matrix and fiber zones, denoted by
, respectively, by employing the relevant constituent
and
damage models. The overall damage factor for both matrix
and fiber zones, denoted by
and , are then evaluated
based on their respective damage methods (i.e. the damage
homogenization for the matrix constituent, and the maximum
damage for fiber constituent, as discussed previously).
Fig. 4. (a) Multi-linear model for the matrix constituent damage, and (b)
linear model for the fiber constituent damage.
After estimation of damage status, the tow effective

material properties in the meso unit cell model were reevaluated from the micro unit cell and the matrix material
properties were degraded. Afterwards, the damage
propagation inside the tows and the pure matrix was
evaluated. For the matrix constituent, following constitutive
relation was used to degrade the stiffness:
(9)
where
denotes the stiffness of matrix zone in the micro
unit cell.
Fig. 5. Damage estimation of the matrix constituent: (a) applied load; (b)
micro stresses; (c) element damage factor; (d) overall damage factor.
Fig. 6. Algorithm combining the MMF methodology and progressive

damage model.
For the fiber constituent, the maximum stress failure

, was used for the fiber
criterion, which is
constituent. The symbol
presents the micro longitudinal
stress of the fiber constituent, and symbols
and
denotes
the longitudinal tensile strength and longitudinal compressive
strength of the fiber constituent, respectively. Fig. 4b
illustrates that when the fiber constituent fails its stiffness is
dramatically degraded. Since, carbon fiber behaves in a brittle
manner, therefore, the highest value of all elemental damage
factors was considered, as following:
Based on the status of overall damage factor, the stiffness

properties of the matrix and fiber are degraded, and the tow
effective properties are evaluated for the next time increment.
The numerical implementation was done using Abaqus 6.11-1,
a commercial finite element solver, combined with its user
subroutine VUSDFLD [9].
52
III. MATERIAL PROPERTIES

The stiffness properties of carbon fiber constituent are
listed in TABLE 1 [1]. However, the strength properties of
carbon fibers, shown in TABLE 1, were back-calculated using
micromechanics approach. The stiffness and strength
properties of the matrix constituent, listed in TABLE 2, were
taken from in-house material testing database.
IV. IMPACT ANALYSIS SETTING

For impact analysis, the whole model of biaxial (BX)
braided fabrics was composed of fiber, pure matrix, matrix in
tows, and a ballistic projectile. For the geometrical modeling
parameters, the tow width and the two thickness used were
3.01 mm and 0.5 mm, respectively. There was a gap of 0.002
mm between the biased tows. The amplitude of the tow
undulation of biaxial braids was set to 0.252 mm. Two kinds
of biaxial braided composites with thermoplastic and
thermosetting resin systems, and each having braided angle of
45 degrees and 25 degrees, hereafter denoted with BX45 and
BX25, respectively, were investigated.
TABLE 1. MATERIAL PROPERTIES OF CARBON FIBERS [1].

Material property
Longitudinal modulus: Ef1 (GPa)
Transverse modulus: Ef2 (GPa)
In-plane shear modulus: Gf,12 (GPa)
Transverse modulus: Gf,23 (GPa)
Major Poissons ratio: vf,12
Major Poissons ratio: vf,23
Longitudinal tensile strength: Tf (MPa)
Longitudinal compressive strength: Cf (MPa)
Proceedings
Value
276.0
27.6
138
7.8
0.3
0.8
3800
2980
TABLE 2. MATERIAL PROPERTIES OF THE MATRIX RESIN.

Material property
Thermosetting Thermoplastic
Elastic modulus: Em (GPa)
3.45
Elastic Poissons ratio: vm
0.35
0.35
65
65
Final tensile strength: Tm (MPa)
Fig. 8. Impact analysis of various types of biaxial braided

composites: (a) braids at 45 degrees, and (b) braids at 25 degrees.
For the ballistic projectile, the radius was 4 mm. The

initial position of the projectile was set above the braided
fabric model and the gap between the bottom of the projectile
and the top surface of the braided models was set to 0.1 mm.
Fixed boundary conditions were applied on the all four sides
of the braided composites, and the top and bottom boundaries
of the braided model were assigned with free boundary
conditions. The ballistic projectile was taken as a rigid body
controlled by a reference point, which was constrained to
move only in z-direction. As illustrated in Fig. 8, the projectile
was given an initial velocity, and the contact properties were
also assigned on the braided model and the projectile by
setting a value of 0.28 for the coefficient of friction. Once the
projectile touches the top surface of the braided model, the
contact behavior will be taken into account. Two initial
velocities of 20 m/s and 50 m/s were applied to the projectile.
At initial velocity of 20 m/s, no penetration occurred and
projectile bounced back after impact. However, at initial
velocity of 50 m/s, the projectile penetrated through the
braided composite.
TABLE 3. TOWS EFFECTIVE MATERIAL PROPERTIES.

Material property
Longitudinal modulus: E1 (GPa)
Transverse modulus: E2 (GPa)
In-plane shear modulus: G12 (GPa)
Transverse modulus: G23 (GPa)
Major Poissons ratio: v12
Major Poissons ratio: v23
Tows with
thermosetting
resin
215.67
13.55
9.0
4.29
0.309
0.592
Tows with
thermoplastic
resin
215.67
13.55
9.0
4.29
0.309
0.592
V. RESULTS AND DISCUSSION

Fig. 9 shows the damage contours for the BX45 braided
composite models, with thermoplastic and thermosetting
resins, at the impact velocity of 50 m/s. For the pure matrix
and the matrix in tows, the damage contour with element
deletion are shown in Fig. 9(a-c) and Fig. 9(b-d), respectively.
Fig. 7. Stress-strain behaviors of the matrix resin systems.
The tows effective properties, listed in TABLE 3, were

computed from micro unit cell, using the stiffness properties
of the matrix and fiber constituents, along with a fiber volume
fraction of 0.78. It should be noted that the total fiber volume
fraction of the meso unit cell was approximately 0.5,
computed in accordance with [2]. The multi-linear damage
models consisting of hardening and softening behaviors,
shown in Fig. 7, for the matrix constituent (i.e. thermoplastic
and thermosetting resin systems) was also established based
on in-house test data.
53
Proceedings
BX25 braided composites subjected to a projectile with impact

velocity of 20 m/s.
(a)
(c)
Dm
Dm
(d)
(b)
Dm
Fig. 9. Damaged matrix element deletion for BX45 braided composite at

impact velocity of 50m/s: (a) pure thermoplastic matrix, (b)
thermoplastic matrix in tows, (c) pure thermosetting matrix, and (d)
thermoset resin in tows.
Dm
Fig. 12. Damage evolution for pure matrix: (a) BX45 braided
composite and impact velocity of 20 m/s, (b) BX25 braided composite
and impact velocity of 20 m/s, (c) BX45 braided composite and impact
velocity of 20 m/s, and (d) BX45 braided composite and impact
velocity of 50 m/s.
(a)
(c)
Dm _ t
Dm _ t
(d)
(b)
Fig. 10. Definition of projectile velocity time histories, before and after
impact.
Dm _ t
Dm _ t
Fig. 13. Damage evolution for the matrix in tows: (a) BX45 braided
velocity of 20 m/s, and (d) BX45 braided composite and impact velocity
of 50 m/s.
In order to get a deeper insight into the impact behavior of

the braided composite models with thermoplastic and
thermosetting resins, the velocity time history of 0.2 seconds
of the projectile was recorded. The velocity before the impact,
is named initial velocity V and the velocity after impact is
named residual velocity V, as illustrated in Fig. 10.
Fig. 11(a-b) shows that for BX45 and BX25 braided
composite models, the impact velocity of 20 m/s of the
projectile decreases with the passage of time. The impact
velocity becomes zero at the time between 0.1-0.15 seconds,
and becomes negative afterwards, representing the rebounding
of the projectile after impact. At lower impact velocity, the
projectile bounces back after hitting the braided composite
model. At time 0.2 second, although the residual velocity of
the projectile for the braided composite model with
thermosetting resin is slightly lower than that of the braided
composite model with thermoplastic resin, however, the
velocity history curves for both resin systems almost overlap,
making it very difficult to differentiate their impact resistance
behavior, at lower impact velocity of 20 m/s.
Fig. 11. Projectile velocity time histories for: (a) BX45 braided
velocity of 20 m/s, and (d) BX45 braided composite and impact velocity
of 50 m/s.
Only the top-side of the damaged braided composite

model, facing the incident projectile, is shown. Similar
damage contours, along with element deletion, were also
observed for the back-side of the braided composite model,
and not shown for brevity. Based on the visual inspection of
the damage contours, it can be qualitatively stated that the use
of thermoplastic resin for the braided composite results in a
better spreading of the impact damage, and lower element
deletion. Similar qualitative observations were made for the
54
On the contrary, the impact resistance of the braided

composite model with thermoplastic and thermosetting resins,
can be clearly distinguished at the projectile impact velocity of
50 m/s, as shown in Fig. 11(c-d). It should be noted that the
projectile with a higher impact velocity of 50 m/s, penetrates
through the braided composite model, as can be seen from the
velocity time histories. The residual velocity of the projectile
decreases with the passage of time, but remains positive
throughout the recorded time of 0.2 seconds. For both BX45
and BX25 braided composite model with thermoplastic resin,
the residual velocity of the projectile is clearly lower than that
of the braided composite model with thermosetting resin. The
decrease in residual velocity of the projectile means that the
impact energy is absorbed by the braided composite model.
Higher the decrease in the residual velocity of the projectile,
the higher would be the impact energy absorption. Based on
this fact, it can be inferred that the braided composite model
with thermoplastic resin exhibits better impact resistance than
that of the braided composite model with thermosetting resin.
resin are compared in TABLE 4. The percentage of element

deletion of the pure resin and of the matrix in tows for the
braided composite with thermoplastic resin system is lower
than that for the braided composite with thermosetting resin
system, as it withstand higher impact loads. Since, the braided
composite with thermoplastic resin system possess better
impact resistance, therefore, the small number of elements
reach the critical damage values that once reach results in the
element deletion.
VI. CONCLUSIONS
A numerical study was carried out to evaluate the impact
resistance of the biaxial (BX) braided composites made of
thermoplastic and thermosetting resin systems, and subjected
to a circular ballistic projectile of radius 4 mm. Two impact
conditions, the bouncing back of projectile with impact
velocity of 20 m/s and the full perpetration of projectile at
impact velocity of 50 m/s, were simulated. Based on the
micromechanics of failure (MMF) theory, the impact
resistance of the braided composite models was evaluated in
terms of the damage evolution in the pure matrix and the
matrix in the tows, and the percentage of element deletion.
Results showed that the braided composites with thermoplastic
resin demonstrate better impact resistance, than that of the
braided composite with thermosetting resin, for both 45
degrees and 25 degrees braiding angles and for all impact
conditions. From the numerical simulations, it was difficult to
evaluate the effect of braiding angles on the impact resistance
of the braided composites, and requires further investigation.
Moreover, future work will involve the validation of the
MMF-based impact analysis methodology, presented here,
with the experimental data.
Fig. 12 and Fig. 13 show that the impact damage evolution

for the pure matrix and for the matrix in tows, respectively, of
the BX45 and BX25 braided composite models with
thermoplastic and thermosetting resin systems. For the pure
resin, it can be seen that the highest damage occurs for both
BX45 and BX25 braided composite models with
thermosetting resin (Fig. 12), for the impact velocities of 20
m/s and 50 m/s.
TABLE 4. PERCENTAGE ELEMENT DELETION.
Impact
velocity
(m/s)
20
50
Percentage
element
deletion
in (%)
Pure matrix
Tow matrix
Pure matrix
Tow matrix
BX45 Braided
composite
Thermoplastic
resin
2.38
0.09
14.0
1.89
Thermosetting
resin
5.38
0.14
20.57
2.04
BX25 Braided
composite
Thermoplastic
resin
3.05
0.18
13.59
1.71
Proceedings
Thermosetting
resin
5.29
0.31
19.36
1.84
ACKNOWLEDGMENT
The authors are thankful to the their colleagues Professor
Dr. Iqbal Hussain and Associate Professor Dr. Aamir Khan of
Mechanical Department at the University of Lahore, Main
campus, 1-kM Raiwind Road, Lahore, Pakistan, for their
timely help and support.
However, in the case of the matrix in tows, the highest

damage occurs for the BX45 and BX25 braided composite
models with thermoplastic resin (Fig. 13), for impact
velocities of 20 m/s and 50 m/s. It can also be clearly seen that
the damage values of the matrix in the tows are higher in
magnitude than that of the pure matrix. Thus, overall damage
status of the braided composite model is governed by the
matrix in tows. Moreover, higher impact damage values also
mean that more absorption of the impact energy and higher
reduction in the residual velocity of the projectile after impact,
as previously illustrated in Fig. 11. Since, damage values of
the matrix in tows for the braided composite model with
thermoplastic resin are higher and well spread (as previous
shown Fig. 9) than that of the braided composite model with
thermosetting resin, therefore, it can be inferred that the
braided composite with thermoplastic resin system can
withstand higher impact loads, thereby, demonstrate better
impact resistance.
REFERENCES
Finally, the percentage element deletion for the braided

composite models with both thermoplastic and thermosetting
55
[1]
Sun, X.S., V.B.C. Tan, and T.E. Tay, Micromechanics-based

progressive failure analysis of fibre-reinforced composites with noniterative element-failure method. Computers and Structures, 2011.
89(11-12): p. 1103-1116.
[2]
Xu, L., et al., Prediction of material properties of biaxial and triaxial

braided textile composites. Journal of Composite Materials, 2012.
46(18): p. 2255-2270.
[3]
Ha, S.K., et al., Micromechanics of Failure for Ultimate Strength

Predictions of Composite Laminates. Journal of Composite Materials,
2010. 44(20): p. 2347-2361.
[4]
Jin, K.-K., et al., Distribution of Micro Stresses and Interfacial Tractions

in Unidirectional Composites. Journal of Composite Materials, 2008.
42(18): p. 1825-1849.
[5]
Raghava, R., R.M. Caddell, and G.S.Y. Yeh, The macroscopic yield
behaviour of polymers. Journal of Materials Science, 1973. 8(2): p. 225232.
[6]
Fish, J. and Q. Yu, Two-scale damage modeling of brittle composites.

Composites Science and Technology, 2001. 61(15): p. 2215-2222.
[7]
Xia, Z., Y. Zhang, and F. Ellyin, A Unified Periodical Boundary

Conditions for Representative Volume Elements of Composites and
Applications. International Journal of Solids and Structures, 2003. 40(8):
p. 1907-1921.
56
Proceedings
[8]
Hashimoto, M., et al., Prediction of tensile strength of discontinuous

carbon fiber/polypropylene composite with fiber orientation distribution.
Composites Part A: Applied Science and Manufacturing, 2012. 43(10):
p. 1791-1799.
[9]
Roget, B. and I. Chopra, Wind-tunnel testing of rotor with individually

controlled trailing-edge flaps for vibration reduction. Journal of Aircraft,
2008. 45(3): p. 868-879.
Proceedings
Improving Physical and Mechanical Properties of

Medium Density Fibreboard (MDF)
Waheed Gul
Afzal Khan, Abdul Shakoor
Mechanical Engineering Department

CECOS University of IT & Emerging Sciences
Peshawar, Pakistan
waheed@cecos.edu.pk
Mechanical Engineering Department

University of Engineering & Technology
Peshawar, Pakistan
Material Preparation
Abstract Physical and mechanical properties play a vital

role in Medium Density Fiberboard (MDF) Manufacturing. Urea
Formaldehyde (UF) resins is easily soluble in water and act as a
binder with fibers. It has fast rate of reaction but has a poor
physical and mechanical properties. Hence, there is a need to
improve those properties of UF resin. The current research is
based to investigates and analyze the UF resin to improve its
physical and mechanical properties in manufacturing of MDF.
The Melamine resin and basic green (Malachite Green) 4-crystals
were introduced into UF resin, separately. In both cases, a
significant improvement in physical and mechanical properties of
UF resin was observed. Additionally, economic analysis was
carried out for both additives and it was concluded that
Melamine is not only economically feasible but also gives
superior mechanical and physical properties when compared to
Malachite green doped resin. The resulted MDF with Melamine
doped resin resulted in strong mechanical adhesion, and water
resistance Keywordscomponent; formatting; style; styling;
insert (key words)
Fiber Separation
Fiber Treatment
Hot Press
Board Trimming
Sanding
Figure 1.1: Production Process flow of MDF [2].
I.
INTRODUCTION
B. Malachite Green
Basic green (Malachite green) is an organic compound,
basically green powder with Metallic luster. The molecular
formula of Malachite green is C52H54N4O12.
Properties of Malachite green are:
It is easily soluble in water
It is extremely soluble in Ethanol.
Its solution in water is Blue-green.
Basic green dyes become yellow in concentrated

sulphuric acid.
The Medium Density Fiberboard (MDF) is a wood product

processed under heat and pressure by using wood fibers or
other plant fibers as raw materials and application of phenol or
Urea-formaldehyde resin as binder. Other suitable additives
may be added to improve the property of the board.[1].The
density of MDF in production is generally control between
690 750 kg/m3.The performance index of MDF is divided
into three categories, i.e. Physical performance, Mechanical
performance and Biological performance.
By nominal density, MDF may be classified in to Type 60, 70,
80. By applicable Conditions, MDF can be classified in to
interior MDF, interior humidity-proof MDF and exterior
MDF.By visual quality and internal bonding strength, MDF
can be classified into Super Grade, Grade A and Grade B. [1].
Greeen Crystals
A. MDF Production Process Flow

MDF production process consists of different stages, i.e.
materials preparation, Fiber separation, Fiber treatment, Mat
forming, Hot pressing, Board trimming and sanding, as shown
in MDF process flow chart in figure.1.
Figure 1.2: Basic Green 4 (Malachite green)
57
II.
Proceedings
LITERATURE REVIEW
Uses of Malachite Green:

Malachite green is used as a dye. It is used as antibacterial. In
Microscopic analysis, it is used as a-biological stain. In dye
laser it is used as a saturated absorber. It is also used as a PH
indicator between PH 0.2 1.8. [5].
Urea Formaldehyde (UF) resins is used as a binder with fibers

in manufacturing of Medium Density Fiberboard (MDF)
because of its fast rate of reaction, easily soluble in water and
also economically feasible. [3].The only problem is that of
poor water resistance property which limits the uses of MDF.
In order to overcome this problem, a melamine formaldehyde
resin and Malachite Green are added with UF resin. If we
compare the market price of both resin, then it can be easily
understood that Melamine formaldehyde resin is almost 4
times expensive than UF resin. So a very small quantity is to
be added for improving water resistance property. The
Melamine resin can easily soluble in both water and UF resin.
[4]. similarly, the market value of Malachite green is pretty
high. Malachite in our case is added to UF resin. The
Malachite is added for the reason to give a unique color and
extra strength to High Density Fiberboard. As we know that it
is easily solvable in water. So the formulation of Malachite
green is carried out in such a way that it dissolves water before
adding it into UF resin. In other words the solution of
Malachite green was added in definite proportion with water
along with caustic.
Malachite green is prepared in two steps: In ist steps
condensation of of benzaldyde and dimethylniline condensate
and leuco Malachite green in obtained. In second step leuco
compound is oxidize to give malachite green. [5].
III.
MATERIALS AND METHODS
UF and Melamine formaldehyde resin and Malachite Urea

Formaldehyde resin have some properties, i.e. PH value, Gel
time, Viscosity, Specific gravity, Solid content etc.
In order to calculate the above properties the following
methods and procedures are used.
A. pH Value Calculation
There are two methods for calculating the PH value of the
Resin.
One way of measuring the PH value is the use of PH paper.
Procedure: The PH strip is immersed in UF resin for a minute.
Then the strip is taken out and compare the pH value with PH
scale (0- 14) numbers. The PH value of UF resin is from 7.5
8.5 means that the UF is alkaline.
UF Resin
65+265(3)265(64(3)2)
2+2 (1)
65(64(3)2)2++122[65(64(
3)2)2]+2 (2)
Manganese dioxide is used as oxidizing agent.
Malachite green can be Hydrolysis and give carbonyl formal.
[65(64(3)2)2]+2
65()(64(3)2)2+ (3)
pH Indicator
Beaker
Table 1: pH values of Malachite Green in Transition
Figure 3.1(a): A PH strip is immersed in Resin
In second method, a PH meter is used to measure the PH

(acidity or alkalinity) of UF. It consists of a glass electrode and
electronic meter as shown in fig. The electrode is connected to
the electronic meter that measure and display the PH value.
58
Flexible Arms
Electrode
Proceedings
C. Gel Time Calculation

The gel time of the resin is the time period in which it
becomes too viscous to flow. The gel time has a very
important role in MDF manufacturing. The gel time
measurement is illustrated below.
Display Screen
Stirrer
Grip Holder
Beaker
Figure 3.1(b): Digital PH Meter
Beaker
Procedure: Before using the Ph meter, the electrode is washed

with distilled water. Then the PH meter is calibrated with
Buffer solution of PH 4, 7 or 10. The UF resin or Melamine
resin or Malachite added resin,s sample is taken in a beaker
and set it on room temperature, i.e. 25 C. The glass electrode
is then dipped to resin sample. The PH value on the display is
allowed to stabilize. After stabilizing the final reading is taken
from display indicator.
Hot Plate
B. Specific Gravity Calculation

The specific gravity of UF, Melamine or Malachite added resin
is measured by an apparatus called Hydrometer. Hydrometer
consists of a graduated tube and a bulb at its end as shown in
fig.3.2.
Procedure: The resin is taken in a graduated cylinder and
adjusted its temperature to 25 C. A hydrometer is inserted in
the graduated cylinder in such a manner that the bulb of the
meter is dipped down. The mercury is raised in the meter and
after stabilizing it on graduated cylinder, the reading is taken.
Figure 3.3: Gel Time Measurement Apparatus
Procedure: Take 50 ml UF resin in a beaker. Now add 1.4 ml

of hardener (Ammonium Chloride) (NH4Cl) to UF resin and
mix it. Now take a sample of mix UF resin in a graduated
cylinder and put it in water bath at 100 C) and on stop watch.
Stirrer the resin till it becomes solid and stops the stop watch.
Now note down the time from stop watch.
Graduatede
Cylinder
D. Viscocity Calculation
Viscosity is the resistance of a liquid to flow. The viscosity of
UF, MUF or Malachite added resin is generally measure with
viscosity cup as shown in fig.5. The viscosity cup has an orifice
of 3mm (ISO standard) at the bottom.
Procedure: A known volume of UF resin is passed through this
orifice and note down the time flow through stop watch. The
viscosity of UF resin is between 200 300 cps (centi poise).
Hydrometer
Figure 3.2: Hydro Meter
59
Proceedings
Vescocity Cup
Pieces of MDF
Stand
Base
Figure 3.6: MDF pieces for Swelling in Thickness
Procedure: Take a piece of MDF specimen 50 mm X 50 mm

and measure its thickness, i.e. T1. Now immersed the sample
of MDF in cold water for 24 hours. The samples are then
taken out from water tray and find its thickness, i.e. T2. The
swelling in thickness (TS) is calculated by the following
formula:
Figure 3.4: Viscosity Cup
E. Solid Content Calculation

The solid content of UF, MUF or Malachite added resin is the
remaining percentage after its drying.
Procedure: take one gram of UF, Melamine or Malachite
added resin and put it in an aluminum foil cup. Now weight it.
This will be Wt. Now put the aluminum foil cup fill with UF
or Melamine resin. in oven at 105 C) for 2 hours. After 2
hours take out the foil cup and find its weight. This will be
W2.
Ts (%) = T1 T2 / T1 x 100
IV.
EXPERIMENTS AND RESULTS
A. UF Resin Properties
UF resin properties were determined using the above
procedures and the properties were then tabulated.
Aluminum Foil
Weighing Scale
S. No
1
2
3
4
5
6
Display Indicator
Table 4.1: Properties of UF Resin

Properties
Values
Color
Transparent/milky white
PH
8.5
Gel Time
100 sec
Solid content
60 %
Specific Gravity
1.26 1.28 kg/m
Viscosity
200 300 cps
B. Melamine Resin Properties

Similarly, Melamine resin properties, i.e. color, PH value, Gel
time, solid content, specific gravity and viscosity were
determined and put in the form of table as illustrated below.
Figure 3.5: Solid content Measurement Apparatus
Then using the following formula find out the solid content:
Solid content (%) = Wt W1 / Wt x 100.
S. No
1
2
3
4
5
6
F. Thickness in Swelling Calculation

Thickness swelling is the difference of initial specimen
thickness to the variation of final thickness after water soak
test in percentage.
Table 4.2: Properties of Melamine Resin

Properties
Value
Color
Transparent
PH
7.0
Gel Time
195 sec
Solid content
66 %
Specific Gravity
1.2 1.24 kg/m
Viscosity
125 150 cps
C. Melamine Urea Formaldehyde Resin Properties

As earlier discussed that UF has poor water resistance. In order
to increase its water resistance properties, 20 gm of Melamine
60
Table 4.6: Constant parameters for 16 mm board
resin was added in 1000 gm of UF resin. The mixture of both

resin were poured into a beaker and shacked it for 5 minutes.
The mixture of resins is then used for determining the
following properties.
S. No
1
3
4
5
6
7
Table 4.3: Properties of Melamine Urea Formaldehyde Resin

S. No
1
2
3
4
5
6
Properties
Color
PH
Gel Time
Solid content
Specific Gravity
Viscosity
Value
Transparent/Milky white
6.7
115 sec
60 %
1.26 kg/m
220 cps
Quantity
80
6
75
Value
9
190
35
16
320
0.9
Units
%
C
Sec
Mm
Sec
%
S.
No
Properties
Value
s
Units
EN
standard
s
Test
Method
Moisture content
7.5
73
EN322
Density
730
kg/m3
720-740
EN323
Swelling in thickness
15
12
EN317
Internal Bonding
0.65
N/mm
0.6
EN319
Modulus of Elasticity
3100
N/mm
2500
EN310
Modulus of Rupture
31
N/mm
30
EN310
Screw Holding face
1015
1000
ASTM1761
Screw Holding edge
741
700
ASTM1761
Table 4.4: Composition of Malachite green solution

Ingredients
Water (H2O)
Malachite green
Caustic (NaoH)
Parameters
Moisture content
Hot press temp
Pres closing time
Thickness of board
Total press cycle
Wax
Table: 4.7 Comparison of 16mm board Properties (UF)
D. Formulation of Melachite Green Crystal

In order to prepare the solution of Malachite green, the
following composition of ingredients were made.
S. No
1
2
3
Proceedings
Unit
Kg
Kg
gm
Samples of the above solution were tests in laboratory for its

PH value. The PH thus calculated through Ph meter was 3.2.
This value indicates that the solution is highly acidic. But as we
know that PH value of UF resin is 8.5.
The Malachite green solution was added in UF resin in such a
proportion that its PH value maintain between 6 and 6.5.
Table 4.5: Composition of Malachite green Urea Formaldehyde Resin
S. No
Properties
Value
1
Color
Blue Green
2
PH
6
3
Gel Time
80 sec
Solid content
61 %
4
Specific
Gravity
1.27
kg/m
5
Viscosity
250
cps
6
For comparison and improved water resistance properties, the

experiments were performed on 16mm boards. One was
manufactured using only 10 % Malachite UF resin and the
other was manufactured using 10 % Melamine UF resin. The
other parameters were kept constant in both cases. The constant
parameters are shown in table below.
Table: 4.8 Comparison of 16mm board properties

With standard value using MUF Resin
S.
No
Properties
Value
s
Units
EN
standard
s
Test Method
Moisture content
8.7
73
EN322
Density
725
kg/m
720-740
EN323
11
12
EN317
Internal Bonding
0.73
N/mm
0.6
EN319
3190
N/mm
2500
EN310
Modulus of Rupture
32.5
N/mm
30
EN310
Screw Holding face
1040
1000
ASTM1761
Screw Holding edge
750
700
ASTM1761
61
Table: 4.9: Comparison of 16mm board properties with standard values

Malachite
S.N
o
Properties
Value
s
Units
EN
standard
s
Test
Method
Moisture content
73
EN322
Density
725
kg/m3
720-740
EN323
12.8
12
EN317
Internal Bonding
0.7
N/mm
0.6
EN319
3150
N/mm
2500
EN310
Modulus of Rupture
31.9
N/mm
30
EN310
Screw Holding face
1001
1000
ASTM1761
Screw Holding edge
739
700
ASTM1761
Proceedings
Melamine consumption /sheet = 0.19 (kg/sheet)

Price of 1 Kg of Melamine = Rs. 150/Kg
Melamine cost per sheet = Rs = 28.5
Similarly,
Malachite Green consumption / sheet = 0.07 (Kg/sheet)
Price of 1 Kg of Malachite Green = Rs = 700 / Kg
Malachite cost per sheet = Rs = 49
Cost Difference = Rs. 20.5
For a 154 M capacity plant, Production/day = 3240 sheets
(16mm)
Cost saving /day = Rs. 66420
So it is clear from economic Analysis that Melamine is
also economically better than Malachite Green.
VI.
CONCLUSION
In the light of the above parameters, samples of 16mm

board were manufactured using the predefined parameters.
The boards were then tested for their physical and mechanical
properties. The tested boards properties were then compared
with the international standards.
The 16mm boards samples using Malachite UF and MUF
resins were manufactured and properties were compared and it
was found that almost all properties were according to the
standards. Moreover, using MUF, the swelling in thickness
property was improved by 15%. And Mechanical adhesion
(which is called internal bond) was increased by 5 %.
Economic Analysis was also carried out and it was analyzed
that Melamine is economically better resin.
Figure 4.1: Comparasion of swelling in thickness of UF,MUF Malachite UF

with standard value
ACKNOWLEDGMENT
The author is highly thankful to Ciel Woodworks Pvt.ltd
Peshawar, Pakistan for giving him the opportunity to perform
experiments in lab and use the factory equipment.
REFERENCES
[1]
[2]
[3]
Figure 4.2: Comparison of internal Bonding of UF, MUF Malachite UF with

standard value
V.
[4]
ECONOMIC ANALYSIS
[5]
As we know from formulation that:
62
ksh, A, and Hosseikhani, H, (2010), Investigation on physical and

mechanical properties of Medium Density Fiberboard (MDF) produced
from hornbeam wood, Volume 25, Number 1 (32); Page(s) 1 To10.
Zwawi Ibrahim, Astimar Abdul Aziz, Ridzuan Ramli, Wan Hassamudin
Wan Hassan and Nahrul Hayawin Zianal (2011), Optimum Parameters
for the Production of MDF using 100 % Oil Palm Trunks, Malaysian
Palm Oil Board, Ministry of Plantation industries and commodities,
Malaysia.ISSN, ppt.1511-7871.
Awang Bono, Yeo Kiam Beng @ Abdul Noor & Kinabalu, Sabah,
(2001),Melamine-Urea-Formaldehyde Resin: changes in Physical
Properties and Strength with Composition Molar Ratio, Borneo Science
10: 11-23.
Jizhi Zhang, Xiaomei, Wang, Shiefeng Zhang, Qiang Gao, Jianzhang Li
,(2013),Effects of Melamine Addition stage on the Performance and
Curing
Behavior
of
Melamine-Urea-Formaldehyde
(MFU)
Resin,,Bejing Forestry University, 100083, Vol 8, No 4.
H. Zare-Hosseinabadi, M. Faezipour1, A. Jahan-Latibari2
andA.Enayati1, J. Agric. Sci. Techno, (2008),Properties of Medium
Density Fiberboard Made from Wet and Dry Stored Bagasse, Vol. 10:
461-470.
[6]
Amee K. Patel, Hardik H. Chaudhary, Khushbu S. Patel and Prof. Dr.

Dhrubo Jyoti Sen,(2014), Colour of Ecofriendly Dyes Used In Holi
Rather Than Triphenylmethane Dyes, Volume 3, Issue9, 1287-1305.
[7]
Nadir Ayrilmis, (2008),Effect of Comparison wood on Dimensions

Stability of Medium Density Fiberboard, University of Istanbul Turkey,
42(2),285-293.
[8]
Cristiane Incio de CamposI, Francisco Antonio Rocco LahrII,

(2004),Production and characterization of MDF using eucalyptus fibers
and castor oil-based polyurethane resin, vol.7 no.3
[9]
Stefan Ganev, (2003),Linear Expansion and thickness Swell of MDF as

a Function of Panel Density andSorptionState,,University
Laval,Quebic,Canada,GIK7P4.
[10]
H. Zare-Hosseinabadi, M. Faezipour1, A. Jahan-Latibari2
andA.Enayati1, J. Agric. Sci. Techno, (2008),Properties of Medium Density
Fiberboard Made from Wet and Dry Stored Bagasse, Vol. 10: 461-470.
[11] Khalid Pervez Bhatti and Muhammad Zuber, (2009), Synthesis and
Application of Melamine Urea Based Precondestates, AUTEX
Research Journal, Vol. 9, No4.
[12] Maintains, G and Berns, J, (2001),Strawboards bonded with urea

formaldehyde resins, 35th International Particleboard Composite
Materials Symposium Proceedings, pp.137-144, 2001.
[13] http://www.goodrichsugar.com/mdfnew.asp?no=3, (April12, 2014.
63
Proceedings
Proceedings
Finite Element Analysis of Tool Wear in Ultrasonically

Assisted Turning
A. Rehman1, S.Maqsood1
R.Muhammad2, N.Ahmed2
Dept.of Industrial Engineering,

University of Engineering and Technology,
Peshawar,Pakistan
Khattak_rehman@yahoo.com
Dept.of Mechanical Engineering

CECOS University of IT and Emerging Sciences
Peshawar,Pakistan
facing greater challenge to machining processes against the

higher
stresses
generation
and
elevated
temperatures[8].Presently, beta titanium is considered as the
most difficult to machine alloys[9].Keeping in view the high
demand of titanium and its alloys machining, researchers have
continuously worked and developed different techniques for
improvement of machinability of these most hard-to-cut
alloys[10]. Such alloys are difficult to machine using
conventional turning(CT) processes and often results in poor
surface finish with associated lack of dimensional accuracy,
built-edge formations, fast tool wear, and undesired chatter
during operations[11].
In recent, new assisted and hybrid machining techniques have
been introduced to get improved machining results of these
hard-to-cut alloys. Among many, ultrasonically assisted
turning
(UAT)
is
one
of
those
in
which
lowervibrationalenergy impacts are imposed on the tool in the
direction of velocity[12-14]. This technique has shown
tremendous and noteworthyenhancement in the machinability
of hard-to-cut alloys [15-18]. In UAT, the tool and workpiece
are not in continuous contact that is why net forces are lower
as compared to CT [12], ultimately friction influence are also
lower [13, 14]. The efficiency of UAT and influence of
vibrational parameters on surface roughness is investigated by
Graeviciuteet. Al[15, 16]. However, limited attention has been
given to investigate the tool life in UAT which is the main
concern in manufacturing processes.
Two of the main critical problems are tool wear and tool
failure, which not only raises the production cost but also
affect the product quality. The thorough and continuous
contact between cutting part of tool and workpiece material
generate high cutting forces and temperature in the cutting
region of hard-to-cut alloys. The selection of appropriate tools
for the machining of these alloys using sustainable
conventional machining processes is still not fully understood
and demands for further research to advance the better tool life
and tool wear.
AbstractThe tremendous increase in use of titanium and its

alloys in different industries, especially in power and aviation
industries demand further investigation due to its exceptional
mechanical and thermal properties. New assisted and hybrid
machining techniques were introduced to obtain better results in
machiningof these alloys. Among many, ultrasonically assisted
turning (UAT) is one of those in which lowervibration energy
impactisdelivered onthe cutting tool in the direction of velocity.
DuringUAT of titanium alloys, momentous improvement has
been noticed in surface roughness smoothness along with
reduction of overall cutting forces, but tool wear behavior is still
unknown. This paper presents a parametric finite element (FE)
analysis for describing tool wear behavior in UAT of (Ti15V3Al3Cr3Sn) alloy on the selected cutting conditions. The
model is useful for optimizing of cutting parameters in UAT for
Ti-15V3Al3Cr3Sn. The proposed FE model is also enhances the
recent experimental work on wear behavior relevant to cemented
carbide tool and comparing UAT and conventional turning (CT)
of Ti-15V3Al3Cr3Sn results.
IndexTermsTool Wear, Ultrasonically Assisted Turning, Ti
Alloys
I. INTRODUCTION
Machining is an imperative practice in manufacturing and
production industries. Machining put in a substantial portion to
the total cost of the product. Machining cost and production
time are the key apprehensions caused by the frequent abrasion
and replacement of machining tool. Compare to milling,
drilling[1, 2]and grinding processes, turning process is
receiving more importance due to enhanced production rate in
advance industries[3, 4].
Uses of Titanium and its alloys are tremendously increasing
day by day in different industries, especially in power,
petroleum, aviation and biomedical industries[5]. In many
applications, these materials replace steels and aluminum
alloys, which usually results in weight and/or space saving,
increase of system efficiency by rising the service temperature,
and remove the need of protective coatings that should be used
in steels [6]. Increment in the use of these alloys is due to its
outstanding mechanical properties including elevated strength
to density ratio, corrosion resistance, fatigue properties and
strength behavior at moderately high temperatures [7].
Aforesaid materials are therefore fall in a category of difficultto-cut materials, because of the fact that these materials are
The subject matter of this work is based on the finite element

simulation of UAT of beta titanium alloy to investigate the
tool life under pre-selected cutting conditions. The simulation
results have been compared with CT. This work would help
the industries to estimate tools life in UAT.
64
II. 3D MODELLING
TABLE.1 Properties of cutting tool
A 3D model for both CT and UAT processes is developed in

DEFORM 3D V 6.1.In current model cutting tool of
Fig.1.
Proceedings
Tool and workpiece FE mesh
Youngs modulus, (Mpa)
560*10-3
Thermal conductivity, (Wm-1C-1)
0.0081*T+11.95
Thermal expansion, (mm mm-1C-1)
9.4*10-6
Heat capacity, (Nmm-2C-1)
0.0003*T+0.57
Poisson's ratio
0.25
A. Cutting Tool
In the present paper, coated carbide insert (AlTiN) is being
used as a tool to machine titanium alloy (Ti-15333). This
coating provides reduction of fraction and adhesion between
chip and cutting tool. In addition to that, AlTiNat high
temperature has good oxidation resistance, wear resistance and
elevatedchemical stability because of the formation of Al2O3
film. Cutting tool mechanical properties are represented in
Table.1. [18]
Tungsten Carbide with (AlTiN) coating is assumed to be

elastoplastic with considering a rake angle as 14.6, nose angle
as55 , nose radius as0.8 mm and cutting-edge radius as0.8
mm. Tool is then meshed with more than 45,000 elements and
is oriented according to the setup proposed. Meshing of tool
and workpiece can be seen in Fig. 1.
Workpiece of Ti-15333is assumed to be plastic and is fully
constrained on lower and lateral side so that it cannot move.
Workpiece used in the simulation is5mm long, fine meshing
been considered with the minimum element size of 0.025 mm
and the number of elements aremore than 30,000. Localized
window meshing has beenselected for the workpiece meshing
to obtain better results. 0.3 mm depth of cut and 0.1
mm/revfeed rate are initially considered for all simulations.
The current model is based on Langrangian methodin which
the chip is formed by continuous remeshing.
B. Workpiece Material
Titanium alloy (Ti-15333) used in our simulation belongs
to the family of beta Ti alloys.These alloys fall in category of
difficult to machine titanium alloys because of the significant
precipitation hardening characteristics. Table.2 and Table.
3[19] shows the composition along with the mechanical
properties of Ti -15333 .
The tool is immovable for the simulation cases of CT, whereas

it vibrates harmonically for the simulations of UAT. The tool
has given an ultrasonic vibration with 20 kHz frequencyand
8m of amplitude in the direction of cutting velocity.Figure. 2
clearly elaborate theschematic of tool and workpiece
movements.
TABLE.2 Ti -15333 (%) chemical composition

Ti
76
V
15
Al
3
Cr
3
Sn
3
TABLE.3 Ti-15333Material properties

Parameter
Unit
value
Solution treated and aged
Fig.2.
Schematic of tool and workpiece movements
65
Density,
(kg/m3)
4900
Youngs modulus, E
(GPa)
87
Tensile strength, UTS
(Mpa)
1200
Thermal conductivity, k
(W/Km)
8.08
Hardness
(Rockwell B)
95
Proceedings
TABLE.4 Cutting conditions details

Experimental conditions
Workpiece Ti-15333 , length 5mm
Cutting tool TiAlN
Cutting speed,(m/min) 10, 20, 30, 40, 50,60
Feed rate, 0.1 mm/rev
Depth of cut, 0.3 mm
Vibration parameters 20 kHz, 8m
Dry cutting
Fig. 3a Stresses-effective in CT
C. Cutting Conditions
Finite element E simulations have been carried out for
overall 12 cases of experiments for both CT and UAT,
consisting of 6 different cutting speeds while keeping depth of
cut and feed rate constant. Table.4 presents simulation matrix
for tool wear analysis.
III. FINITE ELEMENT ANALYSIS
In this section the results of finite element simulations are
conferredand deliberated.Resultant values of stresses and
temperatures are studied which can eventually affected the
tool wear and tool life. Fig. 3a, 3b and 3c shows effective
stresses on UAT and CT processes respectively.We can see
theaverage effective stresses on tool are comparatively lower
in UAT than CTfor one complete vibrational cycle. But on the
other side at penetration stage, high stresses are noted. Fig. 3c
shows the stresses on the tool during retraction time.This
concludes that at penetration stage in UAT, the tool wear rate
is high.
Fig. 3b Stresses-effective in UAT at penetration stage
A. Stresses on tool
In case of CT, the tool is continuouslyin connectionwith the
workpiece whereas in UAT the cutting process can be divided
into four main stages in each single cycle due to the ultrasonic
vibration setup. They are listed as approach, chip contact,
penetration and unloading. In present discussion we are
assuming the tool as at penetration stage and at retraction
stage in each vibration cycle. The number of vibrations cycles
depends on the total simulation time and also on frequency
and amplitude as well. We are considering effective stresses
against the time domain for our analysis of tool wear because
of the ultrasonic vibration cycles and stress fluctuations in
UAT.
Initially, in UAT,when the cutting tool approaches the chip, it
generates smaller effective stresses of 190 MPawhich later on
increases with an increase oftool and chip contact. The higher
Fig. 3c Stresses-effective in UAT retraction stage

stressesof 3140 MPa are noted at the penetration stage of the
tool, which is much more than 3090 Mpa that in CT, see Fig.
3a and 3b. On the other side, at retraction stage, the effective
stresses again reduce to the lower value of 2540 MPa. It can
be concluded that the tool is facing higher stresses at the
66
Proceedings
2500
Stresses-Effective (MPa)
contact point, resulting higher tool wear and reduction in tool

life as comparable to that in case of CT.
The stresses distribution on cutting toolalong the rake face and
the flank face directions in UAT and CT are shown in Fig. 4
B. Tool Temperature
During turning processes an elevated temperature normally
generated in the cutting regiondue tothe plastic deformation
and high friction of tool and workpiece contact. It is also a fact
that high temperature is considered the most important factor
for tool wear and tool life in machining processes. Therefore
evaluation of temperature is an important characteristic for
tool wear and tool life during CT and UAT.
Stresses along the rake face in CT

Stresses along the flank face in UAT
Stressesalong the flank face in CT
Stresses along the rake face in UAT
2000
1500
1000
500
0
0
Comparative analysis of temperature distribution on cutting

tool for both UAT and CT are shownin Fig. 5. The simulation
result enhances the previous work carried out on temperature
effects on cutting tool in CT and UAT. Higher
temperatureshave been noticed at the penetration stage in
UAT.This raising in cutting temperature directly affects the
tool coating regardingthermal softening like reduction in yield
stress,thermal conductivity, Youngs modulus, coefficient of
thermal expansion and specific heat. As the cooling time is
very short and vibration to the tool also increase its
temperature, which further suffice the wear of the tool.
It is also noticed thatif thecutting speedhas increasedfrom 10
m/min to high speed value, the resultant temperature also
increased in UAT. This temperature increment is due to the
decrease in retraction time. The same pattern has been noticed
for other cutting speeds i.e. 30 m/min., 40 m/min, 50 m/min
and 60 m/min. From the above discussions it can be urge that
tool wear is increases in UAT with the temperature increment
at different speed points.
0.5
1
Distance (mm)
1.5
Fig. 4 Effective stresses in cutting and radial direction

in both CT and UAT
Tool Temperature (C) in UAT

Tool Temperature (C) in CT
570
Temperature (C)
500
430
360
290
220
150
80
0
10
20
30
40
50
60
Speed (m/min)
IV. CONCLUSIONS
Fig. 5 Temperature details in CT and UAT
A 3D finite element model has been developed in
DEFORM 3D V 6.1 for the simulation of Ti-15333 in CT and
UAT. The results shows noticeable increase in cutting stresses
in UAT as compare to that in CT. At the penetration stage there
is much higher difference in the effective stresses on the
cutting tool, but at retraction stage considerable lower stresses
are noticed. State variable distributions also show the notable
difference in the effective stresses on cutting edge of the tools
in cases of CT and UAT processes. Further, as increase in
speed from lower to higher value the contact between tool and
workpiece reduces in UAT, resulting high temperatures in
cutting zone, which affected the tool wear and tool life by
increasing the wear in UAT as compared to CT
V. REFERENCES
1.
2.
3.
67
Muhammad, R., et al. Finite-element analysis of

forces in drilling of Ti-alloys at elevated temperature.
in Solid State Phenomena. 2012. Trans Tech Publ.
Muhammad, R., et al., 3D modeling of drilling
process of AISI 1010 steel. Journal of Machining and
Forming Technologies ISSN, 2010. 1947: p. 4369.
Muhammad, R., et al., Thermally enhanced
ultrasonically assisted machining of Ti alloy. CIRP
Journal of Manufacturing Science and Technology,
2014. 7(2): p. 159-167.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
Maurotto, A., et al., Ti Alloy with Enhanced

Machinability in UAT Turning. Metallurgical and
Materials Transactions A, 2014. 45(6): p. 2768-2775.
Arrazola, P.-J., et al., Machinability of titanium alloys
(Ti6Al4V and Ti555. 3). Journal of Materials
Processing Technology, 2009. 209(5): p. 2223-2230.
Veiga, C., J. Davim, and A. Loureiro, Review on
machinability of titanium alloys: the process
perspective. Reviews on Advanced Materials
Science, 2013. 34(2): p. 148-164.
Ezugwu, E., High speed machining of aero-engine
alloys. Journal of the Brazilian society of mechanical
sciences and engineering, 2004. 26(1): p. 1-11.
Dandekar, C.R., Y.C. Shin, and J. Barnes,
Machinability improvement of titanium alloy (Ti
6Al4V) via LAM and hybrid machining.
International Journal of Machine Tools and
Manufacture, 2010. 50(2): p. 174-182.
Zlatin, N. and M. Field, Procedures and precautions
in machining titanium alloys, in Titanium Science
and Technology. 1973, Springer. p. 489-504.
Muhammad, R., et al. Effect of cutting conditions on
temperature generated in drilling process: A FEA
approach. in Advanced Materials Research. 2011.
Trans Tech Publ.
Maurotto, A., et al. Recent developments in
ultrasonically assisted machining of advanced alloys.
in Proceedings of the 4th CIRP International
Conference on High Performance Cutting
(HPC2010). 2010.
Kumabe, J., et al., Ultrasonic superposition vibration
cutting of ceramics. Precision Engineering, 1989.
11(2): p. 71-77.
Muhammad, R., et al., Analysis of a free machining
+ titanium alloy using conventional and
ultrasonically assisted turning. Journal of Materials
Processing Technology, 2014. 214(4): p. 906-915.
Muhammad, R., A. Roy, and V.V. Silberschmidt,
Finite element modelling of conventional and hybrid
oblique turning processes of titanium alloy. Procedia
CIRP, 2013. 8: p. 510-515.
Kim, G.D. and B.G. Loh, Characteristics of elliptical
vibration cutting in micro-V grooving with variations
in the elliptical cutting locus and excitation
frequency. Journal of micromechanics and
microengineering, 2008. 18(2): p. 025002.
Bouchelaghem, H., et al., Wear behaviour of CBN
tool when turning hardened AISI D3 steel.
Mechanika, 2007. 65(3).
Fnides, B., M. Yallese, and H. Aouici, Hard turning
of hot work steel AISI H11: Evaluation of cutting
pressures, resulting force and temperature.
Mechanika, 2008. 4(72): p. 59-73.
Skiedrait, I., et al., Ultrasonic application in turning
process of different types of metals.
19.
68
Proceedings
Maurotto, A., et al., Comparing machinability of Ti15-3-3-3 and Ni-625 alloys in UAT. Procedia CIRP,
2012. 1: p. 330-335.
Proceedings
Metamaterials in Aerospace Industry: Recent

Advances and Prospects
Rabia Zafar, Saima Shabbir
Azeem Zafar
Institute of Space Technology, Islamabad,
National University of Science and Technology, Islamabad,
Pakistan
Pakistan
drsaimashabbir@gmail.com
rabiazafar38@gmail.com
azeemgeologist@gmail.com
Abstract This article offers a review of the existing

metamaterials and their concepts, recent advances, and research
directions. It also describes applications of metamaterials design
in aerospace and defense sector as an emerging technology.
Acoustic metamaterials and metamaterial based antenna present
advantages over conventional noise absorbers and antenna
technology employed in aerospace components. Gain
enhancement of antennas and efficiency of structures are
described in this article with reference to improvements and
advancements made so far in academia. A rational approach to
metamaterials science and engineering can revolutionize
aerospace design and render better prospects for military
applications.
A. Development history
Developmental history of metamaterials began by 20th
century in 1904, when experiments on negative group
velocities were carried out by Pocklington [7] and progressed
through the mid of 20th century contributed by distinguished
scientists [8, 9]. John Pendry in 1990s made significant
contributions in the development of metamaterials theory [6].
Metamaterials do not exist naturally with the exception of
metallic element Bismuth. Plasmas and noble metals show this
behavior in infra red or visible spectrum [10].
KeywordsAcoustics, Aerospace, Metamaterials
B. Existing metamaterials
I. INTRODUCTION
For the past few decades research in metamaterials is
gaining an increasing interest [1, 2] due to counterintuitive
electromagnetic properties incorporated into material
structure. Principally, Metamaterials are left handed
materials which exhibit negative refractive index, unusual
diamagnetic properties, with fractional effective
permeability and a very large in-plane effective permittivity
[3, 4]. The permeability can be attributed to the Eddy
current phenomenon induced in conductive inclusions and
permittivity is attributed to an increased value of dielectric
constant.
For a material to be optically transparent, the incident
wavelength will be refracted off centre the superficial
surface. The angle of refraction is smaller than angle of
incident as demonstrated in Fig. 1. The velocity of incident
wavelength is greater in space than the velocity of
transmitted wavelength [5] and object will appear
transparent.
Metamaterials find immense applications in the field of
science and engineering including wireless communication,
bio sensing, seismology, electromagnetic devices and perfect
lenses [6]. Moreover, metamaterial absorbers and emitters,
super lenses and cloaking devices for military camouflage are
under high consideration.
Metamaterials cover a narrow range of wavelength owing

to the fact that they can respond below red region of visible
wavelength. Based upon the medium properties, metamaterials
can be classified as;
Negative index metamaterials (NIM) which include

double negative (DNG- and are negative) and
single negative metamaterials (SNG-either or is
negative),
Double positive (DPS- and are positive),
Epsilon negative (ENG- is negative and is

positive) and negative (MNG is positive and is
negative). MNGs show magneto optic effect in which
electromagnetic wave propagation takes place in a
quazistatic altered magnetic field [11].
Based on the response to a particular wavelength

metamaterials are further classified as photonic metamaterials,
acoustic
metamaterials,
seismic
metamaterials,
electromagnetic metamaterials, terahertz metamaterials,
tunable metamaterials, elastic metamaterials, chiral
metamaterials and quantum metamaterials.
69
Metamaterial
Proceedings
D. Metamaterials and seismic cloak

The seismic wave guided metamaterials can be made by
stacking sequences of macromolecular polymer based split
ring resonators overlaid with other meta atoms in a
consecutive fashion. Such a metamaterial can be used to
absorb seismic waves generated by earth quake and can
efficiently refract back or absorb seismic waves protecting
building infrastructure.
Incident ray
E. Stealth applications
Conventional
materials
Implementation of metamaterials in stealth technology

finds its use to disguise moving bodies from radar detection;
either by absorbing or by scattering all the radiations from the
radar detection system.
Fig. 1. Ray diagram of a metamaterial and a conventional material
F. Acoustic cloaking
II. MICRO FABRICATION AND APPLICATIONS
Acoustic metamaterials are ordered artificial structures

which can absorb and manipulate sonic, infrasonic and
ultrasonic waves at will [14]. Bulk modulus and mass density
counteract with permittivity and permeability. This analogous
property is under high consideration and research is being
done to direct and control sound waves at will by tailoring
these two parameters. 2D acoustic cloaking has been
optimized by placing five arrays of cylindrical lattices along a
common central axis to achieve omnidirectional cloaking
devices.
Metamaterials industry is emerging as a leading

technology in future for defense and aerospace sectors by
2020. Currently, there are only few methods available for the
fabrication and characterization which involve complex
principles of microelectronics and lithographic deposition.
The characterization of metamaterials is rather expensive,
which poses high need for capital investment in
instrumentation. Due to these limitations there are only a few
metamaterials which show negative effective permeability and
large permittivity and successfully characterized as
metamaterials.
G. Metamaterials antenna gain

Metamaterials are being extensively studied to find
applications in wireless communication and improved antenna
performance i.e. gain enhancement, miniaturization of size,
directivity and bandwidth broadening. Antenna is dispersive in
nature and acts as an electrically small resonator. In a
conventional antenna surface waves propagate to increase
cross polarization effects, low gain, narrow bandwidth and
frequency range.
The quest continues as how to improve the output of
conventional antenna by incorporating metamaterial and
which particular type of metamaterial can improve one or
more parameters? An important consideration to achieve
antenna gain is to incorporate double negative DNG
metamaterials and single negative metamaterials into antenna
structure to harvest efficient yield and enhanced bandwidth.
SNGs are dispersive in nature [15]. Using this principle
behavior SNGs can be alternated with either MNG layer or
ENG layer to get a suitable combination of optical properties,
enhanced dispersion and miniature antenna size.
Finally, to reduce surface wave propagation it is necessary
to minimize the size of antenna and for this purpose photonic
band gap metamaterials can be employed.
A. Polymer based metamaterials

Metamaterials are composed of polymer matrix filled with,
copper, gold or silver flakes. They are formed by stacking
alternating layers of ENGs/MNGs type of metamaterials
separated by space.
Metamaterials for interference shielding applications were
fabricated by pulsed current electrolysis routes and by
carefully controlling the concentration of copper metal flakes
as inclusions in polymer matrix composites [12].
B. Metamaterial coatings
Metamaterials multipole coatings are being applied by
carefully tailoring the surface properties ranging from the
lowest ordered multipole coatings for microscopic bodies to
higher ordered multipoles for macroscopic bodies. Concealing
a macroscopic body becomes more difficult as domain size
increases and concealing the particles in a closed proximity is
difficult to achieve practically.
C. Metamaterial emitters
In a single band metamaterial a noble metal such as gold or
silver is deposited by electron beam lithography on a silicon
substrate followed by curing the spun coated
benzocyclobutene on metallic top. Dual band absorber is
fabricated by lithographic deposition with only difference that
the dielectric space is occupied by Al2O3. Such emitters find
applications in energy harvesting devices [13].
III. Discussion and outlook

The demonstration of recent advancements and analysis of
metamaterials with all possible applications and future
prospects in both academia and industry is interesting. Various
antenna parameters have been studied and improved by
70
different researchers such as gain enhancement, dispersion,

and bandwidth broadening and miniaturizing the size. Studies
revealed that unit cell metamaterials of omega structure can be
combined to form a slab and used to cover antenna surface
and increases directive gain to 11.5 dBi [16].
Metamaterial based broadband antenna were demonstrated
to show -1 dBi antenna gain at frequency of 2.5 GHz. with -1
dBi [17]. Electrically small antenna were demonstrated by
combining layers of ENG metamaterials alternated with
ordered multipole resonant structures to increase radiation
efficiency and high impedance [18].
Metamaterials show exceptional physical properties which
are overwhelming from both theoretical and academic
perspective and definitely find relevance to aerospace and
defense applications.
Proceedings
V. SUMMARY
In this review we have taken a brief outlook of existing
metamaterials, their fabrication routes, current projects and
ongoing research to enhance performance and applicability of
metamaterial devices around the world. Survey of
metamaterials clearly demonstrated that metamaterials can be
used for improving the performance of conventional antennas
and can revolutionize the communication system.
Additionally, a brief review of the current work and
development in the field of metamaterials is provided. We also
examined the on-going projects on metamaterial to enhance
the performance of antennas and various novel metamaterials.
Various future challenges are also considered.
ACKNOWLEDGMENT
Authors would like to acknowledge Department of
Materials Science and Engineering, Institute of Space
Technology, Islamabad for encouragement and moral support
to write this review paper.
A. Metamaterial airborne design

In any purposeful review the electromagnetic properties of
metamaterials, which strongly impinge airborne radome
design cannot be neglected. Overall antenna performance is
strongly dependent on the mechanical and electromagnetic
properties of dielectrics. Unfortunately, there are limited
number of materials which compromise over low dielectric
constant and high mechanical strength. Significant research in
the radome design is necessary in both spheres of
metamaterials design and electromagnetism to circumvent
radome, which may pose restrictions in metamaterials airborne
design in near future.
An outlook of current projects shows that the developed
nations are involved in understanding the metamaterial
concepts and their practical applications. For example, BAE
systems has tested metamaterial based invisible vehicles
which can disguise in infra red and microwave frequency
range. Other international research groups are aiming at
developing nano sized plasmonic devices [19].
U.S. Army Research office in collaboration with university
students is developing liquid crystals tested in the range of 1110GHz. Other tunable devices made by liquid crystals are in
the process of development such as tunable antenna [20].
REFERENCES
[1] A. J. Hoffman, L. Alekseyev, S. S. Howard, K. J.
Franz, D. Wasserman, V. A. Podolskiy, E. E.
Narimanov, D. L. Sivco, and C. Gmach, Negative
refraction in semiconductor metamaterials, Nat.
Mater, vol. 6, pp. 946-950, October 2007.
[2] N. Fang, D. Xi, J. Xu, M. Ambati, W. Srituravanich,
C. Sun, and X. Zhang, Ultrasonic metamaterials
with negative modulus, Nat. Mater, vol. 5, pp. 452456, April 2006.
[3] R. A. Shelby, D. R. Smith, S. Shultz, and S. C.
Nemat-Nasser, "Microwave transmission through a
two-dimensional,
isotropic,
left-handed
metamaterial," Appl. Phys. Lett. vol. 78, pp. 489-491,
January 2001.
[4] D. R. Smith, W. J. Padilla, D. Vier, S. C. NematNasser, and S. Schultz, "Composite Medium with
Simultaneously
Negative
Permeability
and
Permittivity". Phys. Rev. Lett, vol. 84, pp. 41844187, May. 2000.
[5] K. Sanderson, Materials science: Unexpected tricks
of the light, Nature, vol. 446, pp. 364-365, March
2007.
[6] J. B. Pendry, Negative refraction makes perfect
lens, Phys. Rev. Lett., vol. 85, pp. 3966-3969,
October 2000.
[7] H. C. Pocklington, Growth of a Wave-group when
the Group-velocity is Negative, Nature, vol. 71, pp.
607-608, April 1905.
[8] L. I. Mandelshtam, Group velocity in a crystal
lattice, Zh. Eksp. Teor. Fiz, vol. 15, pp. 475-478,
April 1945.
[9] V. G. Veselago, The electrodynamics of substances
with simultaneously negative values of and , Sov.
Phys., Usp, vol.10, pp. 509-514, February 1968.
[10] V. A. Podolskiy, L. Alekseev, and E. E. Narimanov,
Strongly anisotropic media: The THz perspectives
IV. FUTURE CHALLENGES

To address the future challenges metamaterials research
must transcend the theoretical investigations and transform
into practice. Practical applications make transitions through
terahertz (THz), infra red and visible region of
electromagnetic spectrum. THz metamaterials can be tailored
by changing meta atoms at nanoscale and open new horizons
of understanding and research in photonics and electronic
band gap materials. Challenges related to efficiently
generating and detecting broadband THz waves has primarily
limited their use. Currently existing metamaterials respond to
a narrow band of EM spectrum below red region of visible
spectrum. This requires the development of metamaterials to
extend their response in visible, micro and radio wave region
of EM spectrum in both theory and practice. Thus
metamaterials design can miniaturize antenna design [22] and
dispersion coupled with improved satellite communication and
can revolutionize the defense system.
71
of left-handed materials, J. Mod. Opt, vol. 52, pp.

2343-2349, January 2005.
[11] Z. Qiu, and S. D. Bader, Surface Magneto-Optic
Kerr Effect. Characterization of Materials, 2nd ed.,
John Wiley & Sons, 2012, pp.1-10.
[12] P. Los, A. Lukomska, S. Kowalska, R. Jeziorska, P.
Zaprzalski, and J. Krupka, Metamaterials based on
polymer dispersions of nanoparticles and particles of
copper obtained by cathodic current pulse
electrolysis. Materials Science-Poland, vol. 29, pp.
35-40, March 2011.
[13] X. Liu, T. Tyler, T. Starr, A. F. Starr, N. M. Jokerst,
and W. J. Padilla, Taming the Blackbody with
Infrared Metamaterials as Selective Thermal
Emitters, vol. 107, pp. 045901-4, July 2011.
[14] C. Jo, J. Jeong, B. J. Kwon, K. C. Park, and I. K. Oh,
Omnidirectional two-dimensional acoustic cloak by
axisymmetric cylindrical lattices, Wave Motion, vol.
54, , pp. 157-169, April 2015.
[15] G. V. Eleftheriades, and K. G. Balmain, Negativerefraction metamaterials: fundamental principles and
applications. 1st ed., Wiley-IEEE Press. 2005.
[16] A. S. Ahmad, O. Adib, J. J. Mohd, H. B. Noor, A. A.
Robiatun, and F. F. Mohd, Small patch antenna on
omega structure met material, Eur. J. Sci. Res. vol.
43, pp. 527-537, June 2010.
[17] M. Palandoken, A. Grede, and H. Henke, Broadband
Microstrip
Antenna
with
Left-Handed
Metamaterials, IEEE Trans. Antennas Propag, vol.
57, February 2009.
[18] A. Erentok, and R. W. Ziolkowski, Metamaterial
inspired Efficient Electrically Small Antennas, IEEE
Trans. Antennas Propag, vol. 56, pp. 691-707, March
2008.
[19] J. B. Pendry, A. J. Holden, D. J. Robbins, and W. J.
Stewart, Low frequency plasmons in thin wire
structures, J. Phys. Condens. Matter, vol. 10, pp.
4785, June 1998.
[20] D.
R.
Smith,
Research
groupNovel,
electromagnetic materials program, January 2005.
72
Proceedings
Proceedings
Finite Element Simulation of Composite Body Armor

Numan Khan1, Abdul Shakoor1
Dept. of Mechanical Engineering
Peshawar, Pakistan
Write2numan@gmail.com
1
Vadim V. Silberschmidt3
School of Mechanical and Manufacturing Engineering,
Loughborough University,
Loughborough, UK
Riaz Muhammad2, Naseer Ahmed2

2
CECOS University of IT & Emerging Sciences,
Peshawar, Pakistan.
AbstractFinite element simulation of composite body armor
to hemispherical projectile than composite multi layered

target of identical thickness due to free slip between plates.
It was also observed that by increasing the number of plates
in multi layered target plate reduces its resistance, which is
much closed to the data obtained from the experiments of
Almohandes et al. [5]. A detailed analysis of ballistic
performance of composite and hybrid armor plates were
performed using ANSYS AUTODYN software by
Yohannes Regassa et al. [6]. Ceramic faced armor plate
was used for observation of critical penetration velocity,
which shows good agreement with experimental results.
Teng et al. [7] has presented the performance of composite
and single layered steel armor under ballistic conditions and
concluded that nose shape, velocity and mass of projectile
are the key parameters effecting body armor performance.
Similarly, the effect of nose shape and hardness on failure
mechanism of the armor is presented in [8]. Studies of H.
Kurtaran et al [9]. has shown full penetration of
semispherical nose shape projectile through 2mm thick steel
plate at velocity of 500m/s. However the backing of front
plate with a 20 mm thick plate of same material and two
plates at different orientation angle has prevented full
penetration. it was further observed that the results of
Johnson-Cook Material model are more accurate than the
plastic kinematic model for evaluation of the fracture
damage and thermal softening [9]. Various analytical
models have been developed by S. Feli [10]. To analyze the
performance of Alumina plate backed by composite
material made of Twaron fibers. During impact the
fragmentation of ceramic conoid from ceramic tile and
decrease in the semi-angle of ceramic decreases with
increase of initial velocity was absorbed. The research for
design of light weight and efficient body armor required
more efforts for further improvement to increase its safety,
performance and cost effectiveness.
The current study demonstrates the finite element
simulation of projectile piercing ceramic steel target armor.
For fragmentation, the ceramic plate is modeled with
Johnson-Holmquist model where brittle failure, strain rate
effect and enormous deformation are considered. While
backing steel is modeled using Johnson-cook model in
which strain rate hardening effects, thermal softening and
materials fracture are reflected [9].
for impact loading conditions is presented in this paper. The

7.62 mm armor piercing projectile with the velocity of 800 m/s
is used for impacting on the target material. The numerical
model is developed in ANSYS AUTODYN using Johnson
Cook material model for projectile and backing steel material.
The JohnsonHolmquist material model is used for ceramic
material. Different simulations are performed for different
configurations of composite materials and the results of
projectile velocity, kinetic energy, internal energy of target
plate and perforation time is obtained. Finite element
simulation of 6 mm ceramic backed by 2 mm steel plate has
shown full penetration of bullet while the other backed by 4, 6
and 8 mm respectively has prevented complete penetration.
The numerical results show that ceramic material is
responsible for breaking projectile tip and absorbing most of
the kinetic energy. The drop of about 66% in projectile kinetic
energy is observed with increase of backing steel by 2 mm.
Finally optimum ceramic/steel material thickness ratio of 1.5
with light weight is recommended for single body armor.
Keywords- Numerical simulation; Ballistic impact;
Projectile; Armor plate; JohnsonCook material model;
JohnsonHolmquist model
I.
INTRODUCTION
Ever since the discovery of metals and alloys, different type

of protection is made against the enemies attack for
arresting high-velocity projectiles impact by mankind [1, 2].
The development of light weight and effective body armor
is the need of the day due to the current law and order
situation in the world and particularly in Pakistan. For the
mobility of solders light weight efficient and cost effective
body armor is necessary to be developed. Most of the
ballistic materials shows nonlinear deformation and
dynamic behavior under impact loading [3]. Material
properties like strain, strain rate, materials hardening and
thermal softening play a vital role in the design process [2].
With the passage of time different researchers have
developed various ballistic resistive materials.
Normally single high strength metallic plates are used for
body armors. However, multi layered plate configuration
are often used to meet the design specifications which
cannot be fulfilled by metallic single plate. Zukas et al. [4]
studied that a thick uniform steel plate was more resistant
73
II.
FINITE ELEMENT MODELING
Proceedings
Here the
equivalent plastic strain increment and
is the increment in the volumetric compaction strain.
The performance of composite body armor under impact

loading is investigated by explicit finite element model
using ANSYS AUTODYN [11, 12]. To reduce the
computational cost a small circular element of diameter 60
mm is considered for analysis. Initially the thickness of the
target plate is kept constant to 6 mm whereas actual bullet
dimensions are used for bullet modeling in the current
analysis. For addressing the increase of damage effect of
bullet on target plate with decrees in distance, these
simulations are carried out at minimum possible distance.
Similarly for simulation of the actual phenomenon of AK47 bullet with velocity range of 720 to 750 m/s, the
projectile is fired with a velocity of 800 m/s and its
rotational speed is neglected in the current study. Finite
element model for different configurations of ceramic and
steel are modeled as shown in Table 2. The thickness of the
front ceramic material is kept constant to 6 mm while the
4340 steel of different thickness 2, 4, 6 and 8 mm are used
as a backing material. To complete the perforation of the
plate the total time of simulation is considered as 50 s. the
3D model of target armor and 7.62 mm diameter projectile
are shown in the Fig. 1. In current study the simulations are
carried out on core i3 PC with 2.2 GHz processor. The time
taken by one simulation is about between 8 to 12 hours of
CPU time.
Fig. 1. 3D model of projectile and target armor plate
A. Finite element mesh

Fig. 2. The FE mesh details of projectile and target armor plate
A 3-D finite element mesh for both target plate and

projectile is generated by using automatic mesh generation
feature in ANSYS workbench. As can be seen form Fig. 2
the mesh of inner region of plate is kept finer than outer
region. In the mesh the maximum aspect ratio is kept below
five. Both the bullet and plate are meshed with hexagonal
elements of different size between 0.5 to 1.5 mm. The total
number of elements for the projectile and armor are 148,000
including, 52,000 elements for projectile and 96,000 for
armor plate.
Similarly, Johnson-cook material model is used for the steel

part of the armor. The main feature of the Johnson Cook
model is representation[14-16] of the yield strength of
materials on high strain, strain rate, temperature and material
hardening, and is represented by the equation 3 [17]:
(
)(
)( )
(3)
Where A, B, C, n and m are constants and
is representing
value of effective stress, is effective plastic strain,
is
normalized effective plastic strain
The term
can be defined as:
B. Material model
Based on the material characteristics different material
models are used for different materials in bullet impact
simulation [1] In the current study, Johnson-Holmquist
model represented by equation 1 [13] is used for the
ceramic materials. This model is representing relationship
between the intact and fractured strength, pressure and
volume. The various parameter of the model are presented
in Table.1.
[ (
][
)
]
(1)
Where A, B, n, C are material constants and
is
representing the normalized strength of material.
is the normalized pressure. The damage D can be represented
by equation 2.
(4)
In which
and
are used for melting and room
temperatures, respectively. Similarly, the damage of element
in Johnson-cook failure model is given by
(5)
Where
shows the change of equivalent plastic strain and
is the strain at fracture. The term can be represented by
the equation 6.
[
][
] ( )
(
)][
(2)
74
Proceedings
Table. 1 Strength and failure data for SiC using Johnson-Holmquist model
[18]
Parameters
Density
Equation of State Data (Linear)
Bulk Modulus-A1
Strength Data
Modulus of Rigidity
Elastic Limit (Hugoniot)
Intact Strength Constant-S1
Intact Strength Constant-P1
Intact Strength Constant-S2
Intact Strength Constant-P2
Constant of Strain Rate-C
Fracture Strength (Max)-Sfmax
Constant of Failed Strength-
Failure Data
Tensile Limit-T
Constant of Damage- Efmax
Constant of Damage- P3
Constant of Buckling-
Units
(g/cm3)
Values
3.215
GPa
220
GPa
GPa
GPa
GPa
GPa
GPa
None
GPa
None
193
11.7
7.1
2.5
12.2
10
0.009
1.3
0.4
GPa
-0.75
None
GPa
None
1.2
99.75
1
Fig. 3. Variation in projectile velocity
In the same way the Fig .4 gives the dissipation of kinetic

energy for config 1 to 4 after impact of projectile. It is
observed that maximum amount of kinetic energy is
dissipated in ceramic fragmentation and the rest of residual
kinetic energy is observed by the backing steel plate. The
config 4 dissipates maximum amount of energy as compared
to others as it goes through ceramic material. The percent
reduction in kinetic energy is about 66 % with increase of
thickness by 2 mm of backing material. This discloses that
the backing steel plate is responsible for absorption of
residual kinetic energy of the projectile and stop penetration.
The fracture constants are expressed by D1, D2, D3, D4

and D5. For describing the pressure volume relationships
MieGruneisen equation of state is used with combination of
Johnson-cook model [9]. Constant for Johnson-cook model
and MieGruneisen equation of state model are given in
Table 3.
III.
Results of finite element simulation for projectile velocity of

800 m/s and different configurations as shown in Table.2 are
discussed in this section.
The variation in bullet velocity during penetrations for
different configurations is shown in Fig. 3. The config 1, 2
and 3 has maximum residual velocities of about 580 m/s,
230 m/s and 80 m/s respectively after impact while in config
4 projectile has been stopped and velocity is reached to 0 m/s
at 50 s. this shows that the percent reduction in projectile
velocity is about 60% with the increase of backing material
thickness by 2 mm. this reduction is mainly due to the
impact of projectile on ceramic front material.
Fig. 4. Projectile kinetic energy
The variation in internal energy of target armor plate can be

observed form Fig .5 which shows the increase for different
configurations. The Fig. 5a to Fig. 5d reveals that in all cases
the maximum amount of increase in internal energy of
ceramic plate is occurred. It is also observed that the increase
in internal energy of ceramic front plate is about 20 % with
increase of backing plate thickness by 2 mm. Similarly the
maximum increase in average internal energy is observed in
config 4 as shown in the Fig. 5. Thus by increase of backing
Table. 2 Composition and areal density for configuration of ceramic and

4340 steel plate
Configuration
Front
SiC plate
Backing steel
plate
Areal density
(kg/m2)
Config 1
Config 2
Config 3
Config 4
6 mm
6 mm
6 mm
6 mm
2 mm
4 mm
6 mm
8 mm
34.99
50.69
66.39
82.09
75
plate thickness maximum amount of kinetic can be absorbed
Proceedings
in target armor plate.
Table.3 Johnson- Cook model constants for 4340 steel[19]
Density, (g/cm3 )
Elasticity modulus, E (MPa)
7.83
Failure Constants
210000
Strength / Damage constants

A
B
N
c
(MPa)
(MPa)
910
D1
-0.8
586
0.26
0.014
1.03
D2
2.1
D3
-0.5
D4
0.002
D5
0.61
MieGruneisen equation of state constants

0
C1 (m/s)
1.6
3574
Fig. 5. Change Internal Energy for (a) Config 1, (b) Config 2, (c) Config 3, (d) Config 4
76
S1
Ref.
Temp
(K)
1.91
293
increase in thickness as clear form Fig. 7. Another important

point is the deceleration of projectile upon advancement in
ceramic material. This reveals the fact that ceramic material
is contributing mainly in the breaking of projectile tip and
kinetic energy absorption. Finally the areal density for
different configuration is listed. Areal density can found by
adding the multiplication of the density and thickness values
of ceramic and backing steel material [1]. Upon comparison
with steel body armor of thickness 25.4 mm, about 100 kg
per square meter can be reduced. Thus reducing the total
weight can greatly improve the mobility of soldiers.
Config 4
Config 3
Config 2
Config 1
Similarly Fig. 6. shows the bullet penetration for

configuration 1 to 4. The config 1 shows full penetration of
bullet at velocity of 800 m/s whereas in config 2 the
penetration is stopped moderately. However, the bullet is
fully stopped by increasing the backing material thickness as
shown in config 3 and 4 in Fig. 6. Throughout the
simulations the damage mechanisms are observed in the
armor materials.
Firstly, Silicon carbide material which is very brittle tends to
shatter upon impact. It is also observed in the simulations
that the cause of failure of silicon carbide is tensile stress
wave that passes through the ceramic. Secondly, the tip of
the projectile become blunt upon impact with ceramic
material and adopted the shape like mushroom as shown in
Fig. 7. The distortion in projectile shape is increasing with in
Proceedings
t=0 s
t=10 s
t=20 s
t=30 s
Fig. 6. Bullet penetration for different configurations at different time
77
t=40 s
t=50 s
Proceedings
Fig. 7. Bullet shape at 50 s after impact on target armor plate
IV.
CONCLUSIONS
In this work light weight ceramic composite body armor

is analyzed for impact loading of AK 47. A composite
material composed of Silicon Carbide and 4340 steel is
suggested and is simulated for various thicknesses. Based on
the simulation results, maximum reduction in kinetic energy
and increase in internal energy with increase of backing plate
thickness, the final optimum thickness ratio of 1.5 for
ceramic/4340 steel configuration is suggested.
It was also concluded that by reducing the thickness of
the backing materials can significantly reduce the weight of
the body armor, improve strength and will make it easy for
the security forces to carry it. As clear form table. 2 the
weight of about 15 kg per square meter can be saved if 2 mm
less thickness is used for backing material.
[6].
[7].
[8].
[9].
[10].
[11].
V.
ACKNOWLEDGMENT
The authors would like to gratefully acknowledge Dr.

Rizwan Alim Mufti (Head of the Department) Pakistan
Institute of Engineering and Applied Sciences (PIEAS) for
his help in simulations.
[12].
[13].
VI.
[1].
[2].
[3].
[4].
[5].
REFERENCES
Grujicic, M., et al., A computational analysis of the

ballistic performance of light-weight hybrid composite
armors. Applied Surface Science, 2006. 253(2): p. 730745.
Lee, M. and Y. Yoo, Analysis of ceramic/metal armour
systems. International Journal of Impact Engineering,
2001. 25(9): p. 819-829.
Brvik, T., S. Dey, and A. Clausen, Perforation
resistance of five different high-strength steel plates
subjected to small-arms projectiles. International Journal
of Impact Engineering, 2009. 36(7): p. 948-964.
Zukas, J.A. and D.R. Scheffler, Impact effects in
multilayered plates. International Journal of Solids and
Structures, 2001. 38(19): p. 3321-3328.
Almohandes, A.A., M.S. Abdel-Kader, and A.M.
Eleiche, Experimental investigation of the ballistic
resistance of steel-fiberglass reinforced polyester
[14].
[15].
[16].
[17].
[18].
[19].
laminated plates. Composites Part B: Engineering, 1996.

27(5): p. 447-458.
Regassa, Y., G. Likeleh, and R. Uppala, Modeling and
Simulation of Bullet Resistant Composite Body Armor.
Teng, X., T. Wierzbicki, and M. Huang, Ballistic
resistance of double-layered armor plates. International
journal of impact engineering, 2008. 35(8): p. 870-884.
Corran, R., P. Shadbolt, and C. Ruiz, Impact loading of
platesan experimental investigation. International
Journal of Impact Engineering, 1983. 1(1): p. 3-22.
Kurtaran, H., M. Buyuk, and A. Eskandarian, Ballistic
impact simulation of GT model vehicle door using finite
element method. Theoretical and Applied Fracture
Mechanics, 2003. 40(2): p. 113-121.
Feli, S. and M. Asgari, Finite element simulation of
ceramic/composite armor under ballistic impact.
Composites Part B: Engineering, 2011. 42(4): p. 771780.
Fawaz, Z., W. Zheng, and K. Behdinan, Numerical
simulation of normal and oblique ballistic impact on
ceramic composite armours. Composite Structures, 2004.
63(3): p. 387-395.
Sadanandan, S. and J. Hetherington, Characterisation of
ceramic/steel and ceramic/aluminium armours subjected
to oblique impact. International journal of impact
engineering, 1997. 19(9): p. 811-819.
Johnson, G.R., Numerical algorithms and material
models for high-velocity impact computations.
International Journal of Impact Engineering, 2011.
38(6): p. 456-472.
Muhammad, R., et al. Effect of cutting conditions on
temperature generated in drilling process: A FEA
approach. in Advanced Materials Research. 2011. Trans
Tech Publ.
Muhammad, R., et al. Finite-element analysis of forces
in drilling of Ti-alloys at elevated temperature. in Solid
State Phenomena. 2012. Trans Tech Publ.
Muhammad, R., et al., 3D modeling of drilling process
of AISI 1010 steel. Journal of Machining and Forming
Technologies ISSN, 2010. 1947: p. 4369.
Zukas, J.A., High velocity impact dynamics. 1990:
Wiley-Interscience.
Holmquist, T.J. and G.R. Johnson, Response of silicon
carbide to high velocity impact. Journal of applied
physics, 2002. 91(9): p. 5858-5866.
Holmquist, T.J., D.W. Templeton, and K.D. Bishnoi,
Constitutive modeling of aluminum nitride for large
strain, high-strain rate, and high-pressure applications.
International Journal of Impact Engineering, 2001.
25(3): p. 211-231.
78
Proceedings
Incorporate GNSS with Android & Improve the

Search and Rescue Operations
Atiq ur Rehman
Najam Abbas Naqvi
Aeronautics and Astronautics

Institute of Space Technology (IST)
Islamabad, Pakistan
attiqpk2@gmail.com
Aeronautics and Astronautics

Institute of Space Technology (IST)
Islamabad, Pakistan
najm_naqvi@yahoo.com
Abstract Real time positioning or LBS (Location Based

Services) is very important factor that improved the quality of
services thats offered by the institutions in various area,
currently those services are provided using the sophisticated,
expensive and dedicated devices that are used only for specific
applications i.e. LBS or Real-Time Position determinations. In
this research using ICT (integrated Communication
Technologies) devices i.e. Android based Smart Phones and
Tablets/ PDAs (Personal Digital Assistants) that contain sensors
i.e. GNSS receiver, Compass, Gyroscope and Accelerometers etc.
Using open source tools a cost effective solutions is developed for
market. Rescue Operation, Disaster Management needs real-time
positioning services. The GNSS is freely available resource;
currently the GNSS market is growing quickly and penetrates it
in all areas. Now days the GNSS sensors are installed in wrest
watches, phones, tablets and other communications devices.
Android is an open source platform and a famous operating
system of smart device. Freely available resources are used to
integrate both of them android & GNSS to create the application
for population to improve, timely and efficient emergency or
disaster support management.
Open Handset Alliance, a consortium of hardware, software,

and telecommunication companies devoted to advancing open
standards for mobile communication devices.
Android is popular with technology companies which
require a ready-made, low-cost and customizable operating
system for high-tech devices. Android's open nature has
encouraged a large community of developers and enthusiasts to
use the open-source code as a foundation for communitydriven projects, which add new features for advanced users or
bring Android to devices which were officially, released
running other operating systems. The operating system's
success has made it a target for patent litigation as part of the
so-called "smart phone wars" between technology companies.
The android operating system symbol is shown in the fig-1.
Index Terms GNSS, Cellular Communication, Smart

Devices and Android
I. ANDROID
Android is an operating system (OS) based on the Linux
and currently developed by Google corporation. With a user
interface based on direct manipulation, Android is designed
primarily for touch screen devices such as smart phones tablet
PDAs etc, with specialized user interfaces for televisions
(Android TV), cars (Android Auto), and wrist watches
(Android Wear). The OS uses touch inputs that loosely
correspond to real-world actions, like swiping, tapping,
pinching, and reverse pinching to manipulate on-screen
objects, and a virtual keyboard. Despite being primarily
designed for touch-screen input, as regular PCs and other
electronics.
Android's source code is released by Google under
open source licenses, although most Android devices
ultimately ship with a combination of open source and
proprietary software. Initially developed by Android, Inc.,
which Google backed financially and later bought in 2005,
Android was unveiled in 2007 along with the founding of the
Fig-1
II. GNSS
The term Global Navigation Satellite System (GNSS) is
refers to a constellation of satellites that provide signals from
space transmitting positioning and timing data information. A
GNSS provides global coverage. The GNSS receivers
determine location by using the timing and positioning data
encoded in the signals transmitted from satellites. In the world
the 4 types of Navigation systems that are designed and
developed by different countries, the name and their
79
developing countries as The United States of America GNSS is

known as GPS (Global Positioning System) or NAVSTAR.
Russian GNSS is known as GLONASS in Russian Language
known as Globalnaya Navigatsionnaya Sputnikovaya
Sistema is provide the global coverage around the globe as
examples of GNSS. Europe is in the process of launching its
own independent GNSS, system name as Galileo. The Galileo
Navigation system expected to provide the full worldwide
coverage by 2019. The People Republic of China also
developing its own GNSS system known as BeiDou Satellite
Navigation System (BDS) and also known as COMPASS or
BeiDou-2, will be a global satellite navigation system. BeiDou
administration planned to begin serving global customers upon
its completion in 2020.
In this study we discusses the US based navigation system
popularly known as GPS. because the GPS is provide its
coverage around the globe.
Proceedings
satellite constellation consists of 24 space vehicles (Satellites)

(GPS constellation consists of six (6) orbital planes inclined at
55o associated with equator plane, the orbital altitude is
20,500KMs. GLONASS constellation is consists in Three (3)
orbital planes inclined at 64.8o associated with equator plane,
the orbital altitude is 19,100KMs. GALILEIO constellation is
consists in Three (3) orbital planes inclined at 56o associated
with equator plane, the orbital altitude is 23,220KMs. BeiDou
constellation is consists in Three (3) orbital planes inclined at
55.5o associated with equator plane, the orbital altitude is
27,878KMs) The main functions of the Space Segment are to
transmit radio-navigation signals (navigation Data) stored and
retransmit the navigation message provided/uploaded by the
ground Control Segment.
Control Segments:
The Control Segment or Operational Control System, OCS,
is responsible for the proper operations of the GNSS system.
The main tasks performed by the CS are the followings:
Monitor and control of satellites orbital parameters.
Monitor health and the status of the satellite subsystems
i.e. solar arrays, battery power, stabilization etc
Activation of spare satellites.
Regular Update of clock corrections, ephemeris, almanac
and navigation data.
Resolve satellite anomalies.
Controlling Selective Availability (SA) and Anti-Spoofing
(AS) services
Passive tracking of the satellites.
The Control segments comprise of major parts these
areas.
Master Control Station.
Upload Antennas Stations
Monitoring Stations.
User Segment
The User Segment consists of GNSS user terminal that
contain the L-band radio receiver/ processors and antennas
which received GNSS navigation signals, Ephemeris and
Almanac data to determine the accurate pseudoranges. After
receiving the previous mentioned data then solve the
navigation equations in order to obtain their coordinates. A
GNSS receiver is a device capable of determining the user
position, velocity and precise time (PVT) by processing the
signal broadcasted by GNSS satellites.
III. GNSS ARCHITECTURE

The Global Navigation satellite systems (GNSS) consist of
three major parts known as segments. One is The Space
Segment, second is The Control Segment and last one is The
User Segment. These are illustrated in Fig-2.
IV. APPLICATION DEVELOPERS

Android is popular with technology companies which
require a ready-made, low-cost and customizable operating
system for high-tech devices. Android's open nature has
encouraged a large community of developers and enthusiasts to
use the open-source code as a foundation for communitydriven projects, which add new features for advanced users or
bring Android to devices which were officially, released
running other operating systems. The operating system's
success has made it a target for patent litigation as part of the
so-called "smart devices wars" between technology companies.
Android's default user interface is based on direct
manipulation, using touch inputs, that loosely correspond to
real-world actions, like swiping, tapping, pinching, and reverse
pinching to manipulate on-screen objects, and a virtual
keyboard. The response to user input is designed to be
immediate and provides a fluid touch interface, often using the
Fig-2
Space Segment:
The GNSS Space Segment is formed by a satellite
constellation with sufficient satellites to ensure that the users
will have, at least, 4 simultaneous satellites in view from any
point at the Earth surface at any time. The nominal GNSS
80
vibration capabilities of the device to provide haptic feedback

to the user. Internal hardware such as accelerometers,
gyroscopes and proximity sensors are used by some
applications to respond to additional user actions, Android
devices boot to the home screen, the primary navigation and
information point on the device, which is similar to the desktop
found on PCs. Android home screens are typically made up of
app icons and widgets; app icons launch the associated
applications, whereas widgets display live, auto-updating
content such as the weather forecast, the user's email inbox, or
a news ticker directly on the home screen. Third-party apps
available on Google Play and other app stores can extensively
re-theme the home screen, and even mimic the look of other
operating systems, such as Windows or IOS.[59] Most
manufacturers, and some wireless carriers, customized the look
and feel of their Android devices to differentiate themselves
from their competitors.
Various types of Application developer suits are used in
market but Software Development Kit (SDK) is the famous
platform to use for application development in android based
operating devices.
Proceedings
information, see Test your application with the Testing and

Instrumentation framework
App Work Flow

To develop apps for Android, you use a set of tools that are
included in Android Studio. In addition to using the tools from
Android Studio, can also access most of the SDK tools from
the command line. Developing with Android Studio is the
preferred method because it can directly invoke the tools that
need while developing applications.
However we choose to develop with another IDE or a
simple text editor and invoke the tools on the command line or
with scripts. This is a less streamlined way to develop because
you will sometimes have to call command line tools manually,
but you will have access to the same number of features that
you would have in Android Studio. The basic steps for
developing applications (with or without Android Studio) are
shown in Fig-3. The development steps encompass four
development phases, which include:.
Fig-3
D. Publishing
Publishing is the general process that makes your
Android applications available to the users. When you publish
an Android application you perform two main tasks:
Prepare the application for release.
During the preparation step you build a release version of
your application, which users can download and install on their
Android-powered devices.
Usually, release application through an application
marketplace, such as Google Play. However, we can also
release applications by sending them directly to users or by
letting users download them from your own website.
A. Environment Setup
During this phase installed and set up development
environment. And also create Android Virtual Devices (AVDs)
and connect hardware devices on which install the
development applications.
B. Project Setup and Development
During this phase we set up and develop Android Studio
project and application modules, which contain all of the
source code and resource files for designed application.
C. Debugging and Testing
During this phase build an application into a debug
able .apk packages that can install and run on the emulator or
an Android-powered device. Android Studio uses a build
system based on Gradle that provides flexibility, customized
build variants, dependency resolution, and much more. If using
another IDE, build an application using Gradle and install it on
a device using adb. Next, with Android Studio debug
application using the Android Debug Monitor and device log
messages (logact) along with the IntelliJ IDEA intelligent
coding features. We can also use a JDWP-compliant debugger
along with the debugging and logging tools that are provided
with the Android SDK. And an Last, test your application
using various Android SDK testing tools. For more
V. PROPOSED APPLICATION
Few years earlier the firms have no choice to manage &
update changes their area maps by timely surveying. The
rescue response on various emergency conditions response
teams are managed by the area an number of rescue stations are
managed regain, City and country because no one knwn the
batter location of the area maps and totally dependent by the
information that provided by the caller, the caller informed the
condition of emergency and inform his location too, the call
receiver/operator note down information and pointed the
position on given map and then lookout the nearest rescue
operation and mobilized them towards effected areas or parties.
The mention procedure required much time to evacuate the
81
After the installation is done on a device either its a cell

phone or tablet. The application is standby mode because the
power is important factor in the smart devices.
The application is lunched by the App icon at app manager
screen or by dialed fixed or predefined number patters i.e.
1122, 115 & 114 etc. The application starts and performed its
initial operation that is the enable the GNSS receiver, in this
step the physical GNSS receiver is power on and a popup on
the screen the GPS enables.
In the next block is performed operation i.e. tracked the
presented satellites (4 signals) over the sky, and compute its
position, this process take some time up 30 seconds initial for
position fixation. It performed frequently operations until the
position is fixed. In next step, but its performed the parallel
with previous one (position fixing), its desired the information
from the user (depend on the user of application i.e. courier
services, vaccinations, the community health programs or
emergency Search and rescue relief operation) Information
competed with coordinates, now the information is saved at
local storage with automatically generated label (either its date
with time or with number series). The stored information is
may be required to be converted on the required format (i.e.
.nav, .txt or .ext etc extension files or nay other format
according to the requirement) and also added the cell phone
number for easy identification of the position, after suitable
conversion and addition in the file, the information is ready to
send their central information center using the internet or the
SMS, tasks is repeated at three times in four minutes for better
position for rescue operations. Once the task is complete the
user can disconnect the call the application is automatically
disabled the gnss receiver and goes to standby mode.
effected partied from effected place. Since 1992 the first GNSS
is lunched worldwide by Unites States department of defense
knows as Global Positioning System (GPS), this system is
designed for military requirement but the system provides the
navigation information round the clock globally with low
accuracy for civilian users. Initially the GPS receivers huge in
size and quit expensive, Now days the receivers are available
in reduces in sizes and the decreases the price too. The reduced
size receiver are used/ installed in the smart devices i.e. cell
phones, tablets, wrest watches etc.
as on proposed work the main aim is build up an
android based application, thats incorporate the GNSS,
android & Cellular Communication to reduced the time for
rescue/ emergency relief operation by fixing the exact position
of the effected place or parties to save lives and also save them
resources.
A. Application Flow
The android application functional block diagram is shown
in the Fig-4, the functional diagram is operational when the app
is installed on the android based smart device either its a cell
phone or the tablet/PDA, whatever its brand but must have a
GNSS receiver, now days devices have GPS/ GLONASS
supported receivers.
Start
Launched by Dialed
Numbers
Direct Launched
Proceedings
Enable GNSS
Receiver
CONCLUSION
Currently, GNSS resources are available freely without any
cost and GNSS market is growing quickly and has been
expanded in vast areas of applications. Nowadays, GNSS
receivers are installed in wrest watches, phones, tablets,
satellites, aero planes and other communications devices.
Android has an open source code and resource and can easily
be integrated with GNSS. Applications for the local market can
be produced i.e. health, search and rescue will be improved and
enhanced in Pakistan.
Determine Position
with time
No
Position Fixed
Position
Stored
YES
Determine Position
with time
ACKNOWLEDGMENT
Generation
.nav
Extention
Sending
Source
SMS
Internet
Sent to Destination
Server Via Internet
Sent to Destination
Server Via SMS
I personally thank Dr Najam Abbass department of

aeronautics and astronautics institute of space technology
Islamabad, Pakistan for his kind effort for completion of the
research project and express my warm thanks to Mr. Hassan
Ali for his aspiring guidance, and sincerely grateful to them for
sharing their enlightening views on a number of matters related
to the project.
End Process
Standby mode
REFERENCES
Displayed Message
Position Sent
1]
HOANG, V. D., HA, D. T., & PHUONG, X. Q.
(2013). A Vehicle Monitoring and Navigation System Design
based on Android Smartphone (Space, Aeronautical and
Navigational Electronics).
2]
Marti, E., Garcia, J., & Molina, J. M. (2014, July).
Navigation capabilities of mid-cost GNSS/INS vs. smartphone:
Analysis and comparison in urban navigation scenarios. In
Information Fusion (FUSION), 2014 17th International
Conference on (pp. 1-7). IEEE.
Disable GNSS
Receiver
Repeat loop
3 times
Fig-4
82
3]
Djuric, U., Abolmasov, B., Dragana, P., Marjanovic,
M., & Kuzmic, P. PORTABLE GEOTECHNICSUSING
ANDROID SMARTPHONES AND TABLETS FOR
GEOTECHNICAL FIELD INVESTIGATIONS.
4]
developer.android.com
5]
http://en.wikipedia.org/wiki/
Android_(operating_system)
83
Proceedings
Proceedings
Geostatistical Analysis on Seismic Data over

North-Western Regions of Pakistan, Afghanistan
and Eastern Regions of Iran & Tajikistan
Malik Abid Hussain Khokhar1, Asad Ali1, Sadaf Javed1 & Razia Rani2
Department of Space Science, Institute of Space Technology, Islamabad 44000, Pakistan
1
abidmalikgeo@gmail.com, 1asad06@gmail.com, 1nomi4all786@gmail.com,
2
Department of Geography, Government College University, Lahore
2
raziaranigeo@gmail.com
lithospheric masses, grew south of the Pamir that goes to Iran

from touching southwestern parts of Pakistan and the western
regions of Afghanistan. The study area contains a captivating
geological record of Precambrian to present and terminate both
east and west with spectacular bends. The region covers the
foot hills, lesser Himalayas, Greater Himalayas and Higher
Tajikistan Mountainous Zones. Changes in the Earth's
orography and consequent geomorphic changes are directly
related to the ongoing seismic event. This collisional events
have long been argued to be liable for geological, geochemical
and seismic magnitudes of global extent[2]. The exogenic /
eperogenic processes are still going on with the approximate
rate of a few centimeters per year imperceptibly with continued
erosional and denudation factors of internal movements. The
eroded material from its rugged topography is regularly shed
into different depositional settings within the Himalayas. In
this region, the stress and strain caused due to plate movements
which cause frequent earthquakes enormous loss of human life.
Landslides and irregular thrusting have been also cause of
frequent natural disasters. Deforestation and road construction
have also been witnessed as a reason of landslides. Earthquakes
occurred within the Baluchistan Range (Pakistan) and Iran
ranges with the Mw6.4 which generated a complex and poorly
located seismic sequence. These earthquakes mostly affected
Pishin and Ziarat districts along with the borders areas of Iran,
an area where several active faults have previously been
identified. To create the relationship between these earthquakes
variograms models have been established for geostatistical
analysis[3]. In the paper, geostatistical analysis (variogram
estimations, anisotropy, kriging and cross validation) on Depth
and Magnitude of the earthquake data on the subject regions
are applied that display spatial structure ranges and directional
ranges.
Abstract Earthquake, it is a process which usually take place

anywhere in the world. But most of the time this phenomena
takes place at the place which is weak part of the earth - faults.
Earthquakes occur where faults are but magnitude and depth
remain less at particular places. Factually, it is considered that
earthquake activity is normally related to particular activity of
tectonic processes under the earth. The Seismic geomorphic
structures contain particular basic source mechanisms and with
regard to seismic activity, these bases are considered as a solution
of tectonic regional planes. In this way, different seismic activities
solutions bind their strong dependence with a recognized fault
linage system. Furthermore, these connected seismic activities
give us
benefits of attaining an objective of absolute possible
depth and magnitude of earthquake. The seism-o-tectonic
model is an essential parameter that covers the assessment of
earthquake activity. The tectonic framework of the said study
region is extended from the fault system on the west to northwest
Himalaya and boundary of Hindu Kush-Pamir terrain to the
seismic belt of the geographical borders of Pakistan, Afghanistan,
Iran and Tajikistan. Sequel to this, a geostatistical analysis are
conducted by utilizing variogram estimations, anisotropy and
kriging methods for seismic depth conversion and magnitude on
the said sub-surfaces and interval velocity fields in a dependable
way.
Index Terms Geostatistics,
Magnitude, Variogram, Kriging.
Earthquake,
Seismicity,
I. INTRODUCTION
Earthquake take place in the faults areas where their
intensity increases if there is weak surface prevailed then it
bring about the source of a number of earthquakes as a natural
calamity. Various features of geology in the region are direct
manifestation of subsurface of the earth dynamics thats why it
is important to understand variety of problem related to
physical endogenic processes aspects of happening deep down
the earth surface. Earthquakes create faulting and folding due
to above mentioned processes and bring about severe natural
disasters in the affected areas. These happened due to primary
and secondary and long waves that determined depth and
intensity of the earthquakes[1]. The area is extended to
Himalayan arc which encompass about 2,500 km from
northwest to southeast and includes from east to west the width
of the belt varies from 250-350 km. The colossal Himalayas
and the Karakoram, representing the biggest concentration of
II. DATA AND STUDY AREA

Borders of Pakistan, Afghanistan, Iran and Tajikistan are
the study area is chosen where earthquakes are frequent due to
weak subsurface. Area is bounded in between 230.37N to
410N and 440 E to 790 E. The points in the map are showing the
number of earthquakes are frequent due to weak subsurface.
Area is bounded in between 230.37N to 410N and 440 E to 790
E. The points in the map are showing the number of
earthquakes that occurred in between 2014 to March 2015.
84
Proceedings
TABLE I. CALCULATION OF STANDARD ERROR FOR THE SELECTION OF
Sample size of each country is equally taken and total samples

are 150. Major concentration of the earthquakes is the North &
the Western regions of Pakistan and offshore region in the
South, Eastern and the Southern regions of Iran, boarder
regions of Afghanistan, Tajikistan and Pakistan. Magnitude of
these earthquakes is in between 4 to 5.5. Only, Iran is the
country where magnitude is detected 6.6 that occur in the
southern regions of the country.
BEST MODEL FOR PAKISTAN AND TAJIKISTAN

Types of Models
Depth
Magnitude
Linear
1.064598
1.064598
Exponential
1. 064594
1.064600
Spherical
1.064596
1.064596
Gaussian
1.064596
1.064586
TABLE II. CALCULATION OF STANDARD ERROR FOR THE SELECTION OF

BEST MODEL FOR IRAN AND AFGHANISTAN
Types of Models
III. METHODOLOGY
A. UTM Conversion of Coordinates and De-trending Data
Initially, coordinates in the data were in Longitudes and
Latitudes which were converted Universal Transverse
Mercator Projection (UTM) WGS 1984 coordinates by giving
appropriate zones as per the limits of the earmarked countries.
This conversion was necessary to apply other geo-statistical
analysis which are being explicated in the successive pages of
this paper. After UTM conversion, trend was removed from the
data before applying variogram models. Trend is removed from
following equation 1.
2
Magnitude
Linear
47.90590
0.48379274
Exponential
47.03481
0.9713586
Spherical
47.95570
0.5032916
Gaussian
47.78765
0.5004617
Circular
48.00123
0.5035408
For the selection of variograms, exponential model is most

suitable to choose for the results of depths on the regions of
Iran and Afghanistan and this model is selected because of
lowest standard error 47.03481 S.E (Standard Error) among all
the calculated models. Glancing on the table 2, it is evident that
linear model is suitable for the variograms of Magnitude of the
earthquake data as it has lowest standard error 0.48379274.
Figure 2, 3, 4 and 5 are illustrating the variogram fit on the
Depth and Magnitude. Observation on the trend of successive
variograms are depicting cyclical trend where green lines are
fitting lines of the model concerned and red lines in the
variograms are showing estimated models.
Variograms Model.
2
1
(h) m(h) z ( xi ) z ( xi h)
(2)
2
Fig.1. Location of the Study Area
x y I ( x ) I ( y ) I ( x y)
Depth
(1)
Exponential Model.
h
(h) c 1 exp
a
Gaussian Model
2
h
( h) c 1 exp
a
B. Variogram Selection/ Estimation

The variogram or the spatial correlation function describes
the nature/shape and construction of the arbitrary component of
the concerned variable that is the difference between the true
depth of the earthquake and its trend. So, variogram relies
greatly on the trend. The variogram is specified by the standard
error, the correlation range and finally some functional forms
such as Linear, Exponential or Gaussian models. Variogram
models are here to show the correlation of values of each other
i.e. depth and magnitude.
(3)
(4)
Linear Model
(h) c h
a
(5)
Equation 2 which denotes as: - ( h) is variogram, m(h) is

number of point which making pairs, xi + h, xi : variables x at
location i, & i + h: lag vector. Nugget (C0) is the start of
variogram which starts from zero up till its constant figure
which is known as sill (C). Partial Sill is denoted as C1
whereas variogram sills are combination of partial sill and
85
nugget such as C C 0 C1 . Nugget effect is minimized to

zero C(0). Owing to this condition, variogram is not enough for
spatial variation analysis & suitable model must be fitted on
it[4]. In the equation number 2, 3, 4 and 5 are known as the
models as mentioned against each. Here, c is a Sill and h is
lag specified from the distance/ weight derived from the
coordinates and given values at the coordinates.
results.
For
comprehensive
geostatistical
analysis,
omnidirectional variograms are not taken to compute the
directional trend in the data which require to analyze trend of
anisotropy in the sampled data. So sequel to these indicators
directional variograms are calculated with the angle tolerance
150 from 00 to 1650. Figure 6 and 7 are interpreting anisotropic
trends as well as directional trend in the data for depth and
magnitude of seismic data on the subject regions, where pointpaired values are evidently displayed with respect to their
directions Finding enormous trend of direction, it is evident
that direction is not uniform to a specific side rather all sides
which depicts that compactness of the lithosphere which is not
too hard to sustain the shocks of frequent earthquakes thats the
reason for tremendous isotropic trend. Beside these, another
way of determining said trend, there is a method of RoseDiagram that is used for computing anisotropy within the map
having 3600 angular direction, under mentioned figures are
related to said directional trend which are showing the
direction of earthquake magnitude and its depth in the subject
regions. With respect to different angels, an unequivocal
variation in the above shown rose diagrams of Pakistan &
Tajikistan Regions and Iran & Afghanistan regions. When the
data of Magnitude values and depth values of occurred
earthquakes of the subject regions are slanted in a side then
both plots show only that certain direction either 00 degree or
900 degree whereas the trends elucidate the level of depth in
different directions specified rocks of weaken lithosphere. For
comprehending the intensity of the depth and magnitude of the
earthquake, figure 6 and 7 exhibit variogram factor parameters
with all possible angles. Numerical figures of the said
variograms are showing directional variogram for relating
parameters of same dataset within anisotropic influence.
Magnitude and depth have different numbers of values in view
of the sill, partial sill and nuggets and range is comparatively
less differ in the whole dataset because these are based on the
Standard Errors values which are being selected on the basis
of lowest vales and models are also applied on the
appropriations where Depths have 700 at 1500 and sill is 12000
that is the maximum figure in case of Pakistan and Tajikistan
regions and on the same angles Iran and Afghanistan have the
lowest observation at this angle but this regions has more than
6000 sill vales at 1180 and nugget values is 1500. Deviations
proceed to the resultant anisotropic effect in the data values.
Similarly, Magnitude data is of the both regions have different
sill effects, nugget and rang in terms of their direction
significances. Merely, directional variograms are not sufficient
for geostatistical analysis for computations for interpolation
methods like Kriging and Inverse Distance Weighted analysis
but other analysis of variograms are also needed to interpolated
directional ranges and anisotropic corrected variograms of
current data. Above discussed directional ranges are estimated
directional coverage of predicted surface areas along with
appropriate model fittings. Figure 14 is showing Kriging plot
of depth values of earthquake data in all the areas as mentioned
in topic and figure 13 is established using Geostatistical
Analyst Tool of ArcGIS. Figures 12 is serving a clear depiction
of variables Depth and Magnitudes of whole data set with the
C. Kriging
Interpolation provide Kriging methods besides Kernel
Density etc. In geostatistical methods, Kriging is also Best
Linear Unbiased Predictor" (BLUP) which satisfies a certain
criterion and provide minimum variance[5]. The kriging is
based on the weighted moving average[6].
Z x0 i Z xi
Proceedings
(6)
Z It is an estimated var iable at the spatial location

x0 location
Z xi observed parameter of Depth and Magnitude
at every predicted form nearest neighbore of spatial
points
i weights applied to these neighboring points
Unity of sum is the weights are constrained so Equation 6 is

an unbiased estimator[7].
IV. RESULTS
Results in the successive pages are over the seismic data of
the study areas. After the application of various models of
variograms as per the appropriate requirement of model of the
variogram, Ordinary Kriging analysis are applied. It is because
of its real implementation and OK (Ordinary Kriging) is
appropriate applicable where spatial sampled data is large. This
method comprehensively evolves on sampled observations
instead of complete population where trend assessment of
spatial identity always re-assessed so that a smooth and
accommodative model may arise. For depth and magnitude
analysis, Ordinary Kriging (OK) remains always very suitable,
but it does not work properly if we have very small sample data
for using in moving neighborhood observations and so it bring
about the reason of further good and useful results[9]. The
results which showing the variograms are in cyclical with the
models of Exponentials, Linear and Gaussian. These were
derived on the basis of lowest standard that bring about the
enormous appropriate cause of validity of findings for any data
basing upon geostatistical methods. Results are restricted to
bifurcated due to large area and large sample size. Initially,
sample size of the data was accumulatively 600 which split to
two parts of 300 each. Data was full of trend which has been
shown in the variograms models and anisotropy plots. These
plots are derives from the R language and ArcGIS was also
incorporated for visual interpretations which has been useful to
predict the nearby areas by known points. By viewing, figures
12, 13, and 14 are predicting depth and magnitudes of
earthquakes by the application of Kriging and Kriging is found
the most suitable predictors because it is showing very accurate
86
complete area of concentration within UTM coordinates i.e. xlimits 7524 to 1506500 and y-limits 2501400, 4359300; where
filled dots are the known points and the hallow points are
estimated points as estimated with the help of Ordinary Kriging
method of prediction and the same derived from the R
language.. These figures are showing all four areas but with
cut-off extents. If we concentrate on the results, we will be
Proceedings
satisfied by the results of predicted Kriging because of its plots

and cross validation results.
TABLE III. CROSS VALIDATION OF DEPTH AND MAGNITUDE
ON PAKISTAN AND TAJIKISTAN REGION
Pakistan and Tajikistan Regions ( Total 300 Sample Points were Collected )
Cross Validation on Depth
Attributes
Min
Max
Mean
Median
1st Qu
3rd Qu
MPE
MPE/
RMSE
RMSE/STD
RMSE/IQR
MEAN
Observed
1.80
250
64.47
41.55
10
111.5
Predicted
9.13
212.8
65.87
47.33
20.17
109.3
-1.411
-0.2188
26.6608
0.4699597
0.2624747
Residual
-97.4
121.77
-1.4110
0.022
-11.09
9.930
Cross Validation on Magnitude
Observed
5.80
4.456
4.40
4.20
4.60
Predicted
4.32
4.70
4.456
4.369
4.374
4.553
-1.18
-2.6589
0.349549
1
0.9820722
0.8738728
Residual
-0.63
1.369
-0.00118
-0.062
-0.27
0.166
87
Proceedings
TABLE IV. CROSS VALIDATION OF DEPTH AND MAGNITUDE ON IRAN AND AFGHANISTAN REGIONS
Iran And Afghanistan Regions (Total 300 Sample Points were Collected)
Cross Validation on Depth
Attributes
Min
Max
Mean
Median
1st Qu
3rd Qu
MPE
MPE/
RMSE
RMSE/STD
RMSE/IQR
MEAN
Observed
10
265.6
70.32
33
33
98.45
Predicted
9.33
243.9
71.570
38.46
33.15
99.82
-1.252
-0.0178
34.711
0.541
0.5303518
Residual
-155
146.3
-1.253
-1.825
-10.73
3.993
Cross Validation on Magnitude
Observed
6.6
4.537
4.5
4.2
4.7
Predicted
4.23
4.978
4.538
4.523
4.466
4.583
-0.058
-0.0012
0.4161
0.9743606
0.8323063
Residual
-0.94
-0.28
-0.00058
-0.050
-0.286
0.1830
Fig. 2. Exponential Variogram Model on Depth Pakistan and Tajikistan Regions
88
Fig. 3. Exponential Variogram Model on Depth Iran and Afghanistan Regions
Fig.4. Gaussian Variogram Model on Magnitude Pakistan and Tajikistan Regions
89
Proceedings
Fig.5. Linear Variogram Model on Magnitude Iran and Afghanistan Regions
Fig.6. Directional Variograms on Depth Values Pakistan and Tajikistan Regions
90
Proceedings
Fig.7 Directional Variograms on Depth Values Iran and Afghanistan Regions
Fig.8 Interpolated Directional Ranges on Depth Values Pakistan and Tajikistan Regions
91
Proceedings
Fig.9. Interpolated Directional Ranges on Magnitude Values Pakistan & Tajikistan Regions
Fig.10. Anisotropy-Corrected Variograms for Depth - Iran & Afghanistan Regions
92
Proceedings
Fig.11. Anisotropy-Corrected Variograms for Magnitude - Iran & Afghanistan Regions
Fig.12. Area of Plot for Prediction of Depth and Magnitude
93
Proceedings
Fig. 13. Kriging on Depth Values with Depth Unit
Fig.14. Kriging on Magnitude Values with Unit
94
Proceedings
Fig.15. Cross Validation of Depth and Magnitude Predicted Values
Fig.16. Cross Validation of Standard Error
95
Proceedings
Fig.17. Cross validation of Error Values of Depth
Fig.18. Normal QQ Plot
96
Proceedings
Proceedings
Hence, it is concluded that Northeastern regions of

Afghanistan, Eastern regions of Tajikistan are major
concentrated areas of earthquake and more deepened than the
other regions of Pakistan and Iran. However, as for as
magnitudes of the earthquakes are concerned, eastern and eastsouthern parts of Iran are very vulnerable due to weaken parts
of Lithospheric compactness which can give room for more
severe earthquakes with higher intensity of 6.6. Southern parts
of Pakistan with the borders of Iran and western Balochistan
are also vulnerable places for future earthquakes.
V. ROSS VALIDATION
Cross-validation, it is also known as the appreciation with
the name of rotation estimation, that is a technique of model
validation to evaluate geostatistical analysis for generalization
of an independent dataset [10]. It is mainly used in settings
where the goal is prediction, and one wants to estimate how
correctly/ perfectly a predictive model to perform in practice
into implementations of cross validation. It is the tool of
computing the validation of correctness and authentication of
the methodology of geostatistical analysis like Kriging and
Kernel Density etc. It is important to mention over here that
the main reasons for using cross-validation instead of using the
conventional validation is that the error (e.g. Root Mean
Square Error) on the dataset. Because the conventional
validation is not a useful estimator of model performance and
thus the error on the observational dataset does not properly
represent the assessment of model performance[10]. This way
to tool is not only the part of geostatistics methods but also it is
a part of Geostatistical Analyst tool in ArcGIS. Sequel to this,
following are few plots which are derived from Geostatistical
Analyst Tool that includes predicted values, error, and standard
error and QQ plot. QQ plot is showing the distribution of
dataset. Figure 15 is the depiction of distribution of predicted
observations and the sample observations with their crossvalidation.
ACKNOWLEDGEMENT
The authors are indebted to Dr. Asad Ali who guided them
for writing this paper and he reviewed the paper as a whole.
His precious time from his busy schedule and commendable
cooperation is highly appreciable in conducting this research
work.
REFERENCES
VI. CONCLUSION
On the subject topic of four countries (i.e. Pakistan, Iran,
Afghanistan and Tajikistan), six hundred sample points/
observations were collected and only three parameter are used
which are coordinates, magnitudes and depth of the
earthquakes. Geostatistical analysis are applied on these
parameters. With the help of R language, major geostatistical
analysis (variogram model analysis, anisotropic analysis and
Kriging etc.) are conducted; however ArcGIS is also
incorporated for visual interpretation with the utilization of
geostatistical analyst tools. Taking all the analysis methods on
board, it is concluded that Ordinary Kriging is the valid/ best
estimation method because it presents accurate smooth canvass
of surface and minimum standard error estimation as revealed
in cross validation plots and table number 3 and 4. For
correctness of the application of Kriging cross the validation on
depth and magnitude of earthquake data is carried out as well.
Seismic spatial point data are gigantic source of information to
see the location and concentration of earthquake where we can
estimate disaster damage which help us to layout decision
making efforts for benefiting mankind. Semi-variogram and
other variogram models analysis of seismic data are very
useful for estimating destructions and directions of waves of
earthquakes as showing in the figures 2,3,4, 5 and it is also
significant for direction of earthquakes. The paper has shown
that semi-variogram variables of range, nugget and sill which
can be easily used to pinpoint earthquake spatial deviation and
consequently the location of severely damaged areas as well.
97
[1]
N. L. and T. M. Gary Mavko, Seismic Methods For

Imaging Physical Properties of the Earth, pp. 03, 1995.
[2]
S. Development and I. M. Region, The Tajik Pamirs

Challenges of Sustainable Development in an Isolated
Mountain Region. .
[3]
D. L. P. Parasad, Journal of Nepal Geological Society, vol.

45. .
[4]
C. E. G. James, R., Use Of Geostatistics For Accurate

Mapping Of Earthquake Ground Motion, no. 1988, pp. 31
40, 1989.
[5]
V. De Rubeis, P. Tosi, C. Gasparini, and A. Solipaca,

Application of Kriging Technique to Seismic Intensity
Data, vol. 95, no. 2, pp. 540548, 2005.
[6]
M. Franklin, Solution to Ordinary and Universal Kriging

Equations, pp. 15, 2014.
[7]
R. Giraldo, P. Delicado, and J. Mateu, Geostatistics for

Functional Data: An Ordinary Kriging Approach.
[8]
K. Chang, Introduction To Geographic Information Systems,

Sixth Edit. Printed in Singapore: McGraw-Hill, 2012, p.
418.
[9]
P. Abrahamsen, GEOSTATISTICS FOR SEISMIC

DEPTH CONVERSION 3 . Kriging in the geophysical
literature 4 . Seismic depth conversion 5 . Velocity or depth
prediction 6 . Variogram estimation, pp. 19, 1996.
[10]
S. Geisser, Predictive Inference. Chapman and Hall, 1993,

p. 3471.
Proceedings
GIS for estimating optimized water demand using

sustainable water resource management for a planned
city
Rabia Tabassum,1,2, Mudassar Hassan Arsalan2,3, Arif Inam Osmani3, Anam Khalid2
1
Science and Humanities Department

National University of Computer and Emerging Sciences (FAST),
Karachi, Pakistan
2
Geoinformaticsand Sustainable Development Research Lab
Institute of Space and Planetary Astrophysics, University of Karachi, Pakistan
3
Osmani & Company (Pvt.) Ltd., Karachi, Pakistan
.
Abstract
Water, a vital natural resource, is the topic of modern researcher and
getting momentum throughout the world due to its scarcity. Many
regions of the world have reached a point at which present water
resources are already being over-used, as indicated by the depletion
of groundwater aquifer and rivers which no longer to reach the sea.
The alternate water supply approach becomes more important to
achieve sustainable urban development and to reduce the energy as
well. The idea of using modern techniques like remote sensing and
geographical information systems (GIS) in water resource
management studies is quite mature. In the last two decades, RS and
GIS have been broadly used for the preparation of different sorts of
thematic layers and integrating them for multiple purposes,
especially for watershed analysis, ground water recharge and water
demand optimizations.
The DHA City Karachi (DCK), the study area, has been envisioned
as the first sustainable city of Pakistan which will probably serve as
a model for the future independent planned cities in the country and
abroad. The domestic water requirement of 54 gpcd has been
considered as a standard of Karachi Water and Sewerage Board for
the metropolitan city and average water demand per unit area is
calculated as 0.09 g/ft2 for DCK.
This study consists of two parts, in the first part water demand
estimation is done for the DCK land use; Residential, commercial,
Mixed use, Amenities, Recreation, Utilities,
Transportation,
Agriculture farming and Reserved spaces. Water in gallon per capita
per day (gpcd) and gallon per unit area per day (gal/ft2) are
considered as water coefficients for the water demand calculation.
DCK water demand on time scale is estimated as per its development
and inhabitation strategy, which is going to develop in three phases;
Short Term (2012-2015), Mid Term (2015-2020) and Long Term
(2020-2030).
Secondly, to estimate optimized water demand for the land use, the
most feasible and easy to deploy strategies and methods like
efficiency practices (behavior practice and Technology practices)
and gray water recycle and reuse is used. By implementing these
water reduction methods on DCK estimated water demand, it is
evident that recycling and reuse of waste water have a positive
impact on ecological sustainability.
I.
INTRODUCTION
Water is one of the most vital natural resources that somewhat

different from the other, as it is viewed as a key to prosperity
and wealth. Water has played a crucial role in the growth,
function and location of communities[2]. Egypt, South Africa
and Pakistan are the most water-limited nation among the
twenty five most populous countries. According to the World
Bank data, Pakistans river system has a storage capacity of 30
days of water usage, India stores 120 days, while US stores
900 days of river water. In international and national agendas,
sustainable water policies search is at its peak. Approximately
40% of the worlds population will probably experience
shortages in water by 2050. When water could not find in the
desire quality and amount at the right place and time, water
problem may exist [3]. Worlds arid and semiarid regions have
motivated the development of the innovative management
measures due to water shortage. At large, the major part of
Pakistan is arid and semi-arid, therefore water management
system improvement is necessary.
The systems to manipulate and present spatial data can be
considered as Geographical Information systems (GIS),
which are capable of performing a range of function such as
data capture, data storage, data cleaning, data finding and
retrieval, statistical and spatial analysis, and data display [4].
GIS applications as efficacious technology in water resources
management for storing, managing, displaying spatial data are
constantly raising. These applications, including supply and
sewer system modeling, surface hydrology and ground water
modeling, storm water and nonpoint source pollution
modeling for urban and agricultural areas. Because of the
spatial nature of the data for design, planning, operation and
maintenance of water and sewer system in urban areas, GIS
technology used for such systems to provide benefit
significantly. The unique characteristics like family size, lot
size, property value, and soil property, which vary from on
different geographical locations, are associated with
consumers that affect demand[5]. In hydrological modeling,
GIS could be useful to provide the information needed for a
watershed runoff analysis. The integrated technique using
Index Terms:
Water demand efficiency, Recycle and reuse of gray water, Water
demand, Alternate water resources, GIS.
98
GIS and physical based distributed hydrological models

considered as an effective technique in the field of watershed
based water resource management[6].
II.
Proceedings
For the development of a master plan of waste water system,

Giguere [17] used GIS and discovered its efficiency to bring
the relevant data from present databases into geographic
analysis environment, view information and, create report.
The master plan predicts future wastewater flows, assesses
existing treatment and collection system potential, constructs a
sewer rehabilitation program, assesses and identifies
alternatives to expand wastewater facilities, and makes a
phased plan to finance the required improvement. Moutal [18]
presented same work for different country. The automated
mapping system was used to record, update and database
development for planning, design and operations.
Domestic water demand considered as a complex element of
different factors, such as socioeconomic and physical
characteristics, urban planning strategies, infrastructure, and
public water policy. In GIS environment, different thematic
layers are used such as road network distance, distance from
the city center, distance from the coastline, topographic slope,
land use/land cover, Urban Plan, population density and
existing water supply and sewerage system that correspond for
the planning factors.
This study is basically for newly planned city and GIS tools
not only used for the master plan of the city, but also for
estimating water demand on spatial and temporal bases. The
main emphasis of the study is the estimation of optimized
water demand using sustainable water management approach.
LITERATURE REVIEW
In the last few decades, GIS use started to proliferate in the

field water resources management and planning. Mark R.
Leipnik and Hugo A. Loaiciga [7] discussed eachaspects of
GIS and focused on the other facets of GIS pertinent to the
planning and management of water resources. Geographical
Information Systems (GIS) have become an integral and
beneficial tool in many analyses such as statistical and spatial
analysis in water resources management [4, 8, 9].
Panagopoulos, et al. [10] estimated the future water
requirement based on population growth using multi-criteria
decision analysis in which Analytical Hierarchy Process
(AHP) and GIS techniques were involved. AHP has proved as
a useful tool for planners that assess future water demands and
it can be suitable for planning of a national water management
system.
Udovyk [11] s research revealed that GIS couple with
modeling have potential to make a significant contribution to
integrated water resources management (IWRM) and also
notified the issue involved in using GIS as part of IWRM. GIS
considered as a powerful tool to interpolate and model the
spatial data, enable to overlay and compare the environment
and population data for quantifying the population at risk, and
also communicate these to decision makers and other
stakeholders[11].
A particular aspect of water resource management problems,
needs a specific approach to the development of GIS data
structure. By constructing a conceptual GIS data model to
integrate the physical and logical components of the modeling
problem into an operational framework, Daene C. Mc Kinney
[12] expended the function of GIS that implements a tight
linkage between Water resources management and GIS model
Ahmad [13] introduced a new approach for Simulation of
Water Resources Systems called spatial system dynamics
(SSD which presented for modeling feedback dynamics
processes in time and space. GIS and System dynamics are
couple to develop this model. In the study presented by Allen,
et al. [14], GIS is used for verifying and calibration a
continuous simulation model with an event driven model;
they realized that GIS is useful in enhancing the modeling
efforts because of increased accuracy, and minimization of
human error with time.
One of the GIS application is future forecasting of water
demand, Shamsi [15] presented this application
and
established the criteria for future development with the help of
potential development map, a digital elevation model, a
zoning map, and water ,sewer and road network maps. Same
works could be made for sewerage flows for the rapidly
growing population areas. For calculating and projecting the
average daily sewerage flows from the different land use such
as residential, commercial, and industrial, Kirshen [16] used
data handling properties of GIS.
III.
STUDY AREA
DHA City Karachi (DCK) has been envisioned as The First

Sustainable City of Pakistan, which will serve as a model for
the future independent cities in the country and abroad. It is
with the aim to provide its citizens a safe, comfortable,
efficient, and beautiful lifestyle. DCK is strategically located
on the Karachi-Hyderabad Super Highway in District Malir at
the eastern border of Karachi as an independent urban center
in the vicinity of Karachi Metropolis. The city is planned on
an area spanning 11,668 Acres, comprising residential,
commercial and mixed-use elements.
In geographical characteristics topography and distances are
the main factors affecting the potential water sources. DCK is
naturally drained by Abdar and SukkanNullah (small nonperennial water tributaries of Malir River). Their productive
aquifer is recharged from the same water streams and small
water reservoirs developed on them. It extends from latitudes
240 57 N to 250 2 N and longitudes 670 24 E to 670 32
E, and is a sub-tropical location in character.
The study area has a moderate climate and experiences bulk
rainfall in the duration of monsoon season, but the average
rainfall is recorded as only 7inches per annum. In summer,
May and June are the hottest months and temperature often
reaches the 43C. January is considered as the coldest month
with temperature 5C. As a part of Karachi district, DCK may
get its share of water from the bulk water supply by Karachi
99
Water and Sewage Board. But the current water budget of KW

&SBs has been already in stress and demand is more than
supply.
Proceedings
shape file. The master plan of DCK was consisting of detailed

parcels/plots, transportation corridors, reserved spaces for
natural drainage, open and green spaces etc. Land use
classification was applied to parcels / plots, according to the
detailed master plan categories on different levels, such as the
major level comprised Residential, commercial, mixed use,
amenities, recreation, utilities, transportation, agricultural and
reserved spaces. The minor / detailed land use maps were also
developed for each individual major land use class according
to DCK planning. The flow chart of land use layer formation
is shown in Figure 3.
Figure 1: Study Area DHA City Karachi with its land use
A.
DCK land Use Planning
The DHA City Karachi is the study area of this research and
Figure 1 shows its land use. For DHA city Karachi the entire
planning approach focuses on Sustainability Design
Principles, including the concepts of connectivity, efficiency,
renewability, minimization of externalities and carry
capacity[19]. These goals of sustainable design have been
adopted in the overall districts planning approach by
maintaining the ecological integrity of site nullahs, reserved
spaces and natural contours[20]. Master plan of DCK
cumulatively covers approximately 11,668 Acres of land and
comprises of eight major land use types as shown in Figure 2.
Figure 3: Flow chart of land use layers formation
The water demand for DCK was calculated in the spatial

domain, on the basis of water in gallon per capita per day
(gpcd) and gallon per unit area per day (gal/ft2), which was
measured according to the land use types. On a time scale
DCK water demand is estimated as per its development and
inhabitation strategy, which is going to develop in three
phases; Short-term, Midterm and Long-term. Sewage volume
is calculated, which gives the amount of water that can be
reused for the non-potable use. Also by implementing the
water reduction method such as efficiency practices and grey
water recycling, optimized water demand is estimated.
Figure 2: Land use types and % of Area of DCK
IV.
METHODOLOGY
The basic aim of this study is the estimation of optimized

water demand by using alternate water resources for DCK in
GIS environment. For this purpose, the master plan of DCK
was converted into GIS format. ArcGIS software was used to
translate files from AutoCAD drawing files to the ArcGIS
A.
Criteria of Water demand estimation
According to United Nation, in 2012 approximately 50% of
the world's population lived in urban areas and this percentage
is expected to swell to 60 % by 2030[21]. Therefore, water
demand in commercial, domestic, and industrial areas are
consequently growing in cities and result the water
scarcity[22]. For newly planned city, water supply authorities
100
are needed to supply water to their consumers with a certain

degree of reliability. Non uniform distribution of water will
cause water scarcity in some area and excess in other areas
which result the obstacle the natural growth of the city.
Therefore, proper criteria setting is necessary to know the
water demand in spatial and temporal domain instead of
computing demand as a whole which is the main objective of
this study.
Municipal water use commonly consists of residential,
industrial, commercial, and public uses, as well as some minor
use for other purposes like system losses, line cleaning and
firefighting. In developing countries public water withdrawal
is only 13-26 gpcd in contrast, many developed countries
water withdrawal is 79-158gpcd. Municipal water demand is
exactly related to the amount of water withdrawn by
population in town, cities, housing estates, public and,
domestic service enterprises.
The public demand consumes high quality water from the city
water supply system and as its supply also includes water for
industry that provides directly for the requirements of urban
populations. The portion of municipal water demand that is
used in the home is the residential or domestic water demand
and worldwide typical domestic use of water in comparison
with residential plot of 500 square yards is given in Table 1.
Table 1: World Wide Typical domestic use of water in
comparison with residential plot.
World Average
(%)
Types of Water Use
Residential
500 sq.yd plot
(%)
Toilet Flush, including its leakage
33
30
Shower and bath
28
25
Wash basin
11
10
Kitchen
10
Drinking, cooking
(2 - 6)
Dishwashing
(3 - 5)
Garbage Disposal
(0 - 6)
Proceedings
Estimation of water demand based on per-capita consumption

is one of the common approaches. Karachi Water and
Sewerage Board is the main source of water supply for
Karachi city. DCKs current estimated water demand stands at
45MGD. Initially the average domestic water requirement has
assumed as 54 gallons per capita per day for the DCK which is
a reasonable value in comparison with per capita daily water
demand for the different world cities, e.g. Tokyo Japan
(77gpcd) New York-USA (161 gpcd) Colombo-Sri Lanka
(60gpcd) and Mumbai- India (43gpcd) [24].
Most of the people who have purchased plots of land for
residential or commercial purposes in the DCK are likely to
belong to the affluent society of Pakistan. The houses or
dwellings constructed by them are considered to be of superior
quality depicting a higher standard of living, which requires a
large amount of per capita water consumption. It is anticipated
that this may necessitate a higher per capita water supply to be
provided for the water supply system in this area. In
comparison to the general average provision normally adopted
by KWSB for their water supply project as for Karachi city,
which is only 36gpcd[25].
Commercial, Industrial, and Institutional (CII) sectors are
considered as the significant contributors to Public water
demand. Historically utilities have relied on water use
coefficient, which is the number of employees for estimating
CII water use, but it is difficult to obtain this information at
affine resolution. To overcome this situation through spatial,
physical and economic property-based information, Miguel A.
Morales [26] developed a methodology to estimate CII water
use. For calculating commercial water demand, a survey has
been conducted in different commercial areas of Karachi.
With the help of a questionnaire, commercial water demand
has been estimated on the basis of water uses for drinking,
cleaning, washroom and other[27]. The result shows that
water demand per employee has been recorded as 2.3 gal per
employee per day, whereas the water demand per unit area
was 0.031gal/ ft2. Generally, the commercial water demand is
about 10-20% of total water demand. The average water
demand as per planning of the DCK commercial plot is given
in Figure 4.
Laundry and washing machine

16
15
Lawn
sprinkling
and
miscellaneous
3
10
*Source: Interim Report-1 feasibility report of DCK alternate water supply
[1]
This water use is highly dependent on socioeconomic and

environmental conditions and including toilet flush, bathing,
cooking, drinking, watering lawn, and other uses. The average
water demand varies from 25 to 120 gpcd whereas the most
commonly used range is 50-80gpcd for the world urban areas.
In many developed countries residential water use can
constitute over half of total municipal water use and
consumption. Residential water use is also directly linked to
the general public health. Improving the health of the poor is
one of the main goals of water and sanitation projects.
Therefore, household water use is always the top priority in
the municipal water supply[23].
Figure 4 : Average Water Demand as per planning of DCK

commercial plot.
101
Water use in public building such as community hall,

Marriage hall, Mess club, culture and art, school etc. as well
as in public services including fire station, street washing and
landscape irrigation are considered as a Public water use.
Usually 5-10% of Municipal water demand use in public
area[24]. As per planning of DCK the average water demand
for different public building are given in Table 2. The public
water demand may become more precise when the actual
detailed building structure is planned.
Proceedings
In the whole water requirement, the biggest share is of

residential use (41% of total water demand) which occupies
the major share of land in DCK (37% of land use). However
the water demand per square feet for residential area is quite
low (0.09g/ft2). The further distribution of water requirement
on various sizes of plot in the DCK residential area is shown
in Figure 5. It shows that approximately 46% water will be
required for R-5 (i.e. plots of 500 square yards).
Table 2 Average water demand for public water use in DCK

Land Use
Community Hall,
Hall,
Average Water Requirement

Marrige
0.15 gallon per sqft per day
Mess , Club, Culture & Art
Mosque
Hospital
250 gallon per bed

0.15 gallon per sqft per day of
covered area
DCK Clinics
Collage/ School(day scholar,
without boarding)
17 gallon per student per day
University
34 gallon per student per day
Glof club
0.02 gallon pers qft per day
Lake View Park
Park and Playground
Theme Park
Nursery
Road Landscaping

V.
Figure 5: Distribution of Water demand (MGD) in

Residential Land use
In the DCK almost 2% area is designated for commercial plots
of various sizes in community centers and in the central
business district. However, some of the largest plots with
higher FAR are planned as the Mall Zone in a commercial
area and has the highest water demand as compared with the
other commercial areas as shown in Figure 6.
RESULTS
A.
Water Demand for DCK
A key input in municipal water services planning and design is
the estimation of present water demand, and the prediction of
future water demand. In distribution of domestic water equally
to all regions with proper pressure, a uniform spatial water
distribution system will be very helpful[28]. As the total
average water use when DCK will be fully developed
expected to reach up to 45MGD and therefore water demand
per unit area is around 0.09g/ft2. Land use planning is used for
estimation of water demand for DCK. Table 3 shows the
distribution of water requirement for the land use of DCK.
Table 3: Average Water demand of fully developed DCK
S- No
Land use Types
Water
Demand(MGD)
Figure 6: Distribution of water requirement (MGD) in

Commercial Land use.
Water
Demand %
Residential
18.6
41.2
Commercial
7.6
16.8
Mixed use(Res. Cum Com)
7.4
16.4
Amenities
3.7
8.3
Recreation
0.7
1.7
Utilities
5.4
12.0
Transportation
1.0
2.3
Agricultural and Reserved

Spaces
0.6
1.2
Total
45.0
100.0
Its vertical height (FAR) and horizontal size (plot size) with
density of water users (population) are the major water
demand controlling factors. In the overall DCK the second
highest water requirement will be for commercial plots, which
is almost 16.8% of the total water demand. Although most of
the mixed use plots in community centers (C3, C4 and C5) are
comparatively small in size, however large designated plots in
the CBD and South Zone are the major role players due to
their sized and respective FAR. Further distribution of water
demand for mixed use plots is given in map of Figure 8. It has
102
been observed that the water need for mixed use plots is
16.4% of the total water demand of DCK and almost same as
for commercial plots.
Proceedings
There are some reserved areas within DCK which are required
to be kept green that includes right of way (ROW) of oil and
gas pipelines passing through the DCK reserved areas for
agriculture and some open and green spaces along the storm
water drainage corridors. These areas will require around
1.2% of the total demand with low water requirement per unit
area (approximately 0.01g/ft2). The water distribution map of
these areas is given in Figure 9.
Figure 8: Distribution of Water requirement in Mixed use
A long list of required amenities and special used is planned in

DCK that covers around 7.3% of DCK land. It consists of
Educational institutions, public building, religious center,
health facilities and some reserved amenities for future
requirements. The average water requirement per unit area for
amenities is 0.16 g/ft2 with an overall requirement of 8.3% of
total water demand. Educational institutions are the main
water demanding units due to the major strength of water
users (students). Further detail of average water requirement is
given in map of Figure 7.
Figure 9: Distribution of Water requirement in

Agriculture and Reserved Spaces
As a first sustainable city of Pakistan, DCK is planned in a
great comprehensive manner that is going to minimize its
dependencies up to the highest extent on the resources outside.
In this connection its entire energy requirement is planned to
produce at the site with the blend of renewable and
conventional energy sources. Similarly, the water requirement
is going to be fulfilled by optimizing the conjunctive use of
alternate water supply with regular water supply from the
bulk water source.
Almost 3.3% of the land have been designated for exclusive
use of utilities to manage in a sustainable manner. The overall
average daily water requirement for the complete city is
around 12%, with the lion's share for various processes and in
conventional power generation (i.e. 79% of the total utility
requirement). Its per unit area requirement is quite higher
(0.34 g/ft2). Further details are given in map of Figure 11.
Figure 7: Distribution of Water requirement in Planned

Amenities and Special Use
The main requirement of water in recreational use is related to
the landscaping and green space development. The estimated
water demand coefficient is 0.1 g/ft2for recreational areas.
Approximately 1.7% of DCK land is designate for recreational
purpose. This share will not only serve to the local
recreational demand, but also cater for regional requirements
of high quality amusement and sport facilities such as theme
Parks, golf course and specialized sport centers as shown in
Figure 10.
Figure 11: Distribution of Water requirement for Utilities

In the whole DCK urban Green is in focus. With good
architectural design, landscaping is planned with multi-
Figure 10: Distribution of Water requirement in Recreation
103
2)
Figure 13: Distribution of Water requirement in

Transportation use
parameter. Fruits, flowers, flower colors, fragrance, birds
attraction and water requirement are the main selection
parameter for soft landscaping. In this connection other than
recreational facilities, transpiration corridors are also planned
with soft landscaping. Transportation corridors cover around
34% of DCK land with different row and road categories. The
water requirement for transportation corridors landscaping and
other minor uses will be around 2.3% of total water demand
with highest share at streets. The details are given in map of
Figure 13.
1)
Spatial Analysis of Water Demand
According to zonal distribution, DCK comprises of four main

zones; DCK sectors, South Zone, Down Town and Gateway,
as well as miscellaneous area as shown in Figure 14. In zonal
distribution more than half (60%) of water share is required
for residential sectors with a comparatively low water
requirement per unit (0.07g/ft2). Second largest requirement of
water is in south zone that is almost 15.2 % of the total
Proceedings
Temporal Analysis of Water Demand
Many of engineering applications and operational

management of water distribution system are based on the
future water consumption. Therefore, it is necessary to
forecast water demand not only on spatial scale but also with
development time considerations. It is very hard to get a
reliable estimation of future water demand without a clear
understanding of historical and current trends of water use
patterns [29]. Commonly, water demand prediction is divided
into three categories: long - term, medium term and shortterm periods. Water forecasting in future years used for
distribution system design refers to long-term period. Seasonal
water consumption refers as median-term period, whereas
short term period concern with daily and hourly water
demand.
Domestic water use by households is a dominant municipal
water requirement in urban areas. Water utilities to make
system plan, design and asset management are important for
municipal water demand, short- and long- term forecasting.
Long term water demand is useful for management and
operation, whereas short term is essential for municipal water
demand short and long term forecasting. Water demand
prediction accuracy achievement is challenging, however ,
because the predicting model must simultaneously consider
multiple factors associated with economic development,
population growth and migration, climate changes, and even
consumer behavioral patterns[30].
The population growth rate of the city is required for
forecasting the domestic water requirements. Future
population prediction is based on the knowledge about the
Table 4: Water requirement with respect to DCK development
phases
Development plants
Water requirement
after completion
(MG)
Yearly Water
Resource
Development
requirement(MG)
Short-Term (2012-2015)
11.1
3.7
Mid-Term (2015-2020)
20.7
4.14
Long-Term (2020-2030)
11.5
1.15
Figure 14: Zonal Wise water demand
demand with a comparatively higher value per unit area

(0.15g/ft2). The highest water requirement per unit area in City
Gateway and Downtown (i.e. 0.27g/ft2), that covers around
14% share. Further sectors wise water requirement are given
in Figure 12 and Figure 15 that show the sector wise split of
land use water demand.
Figure 15: Sectors wise water demand of DCK
city, its development of surrounding areas, environments,

potential trade expansion and, infrastructure and
communication network development. As per DCK
development strategy, it is going to develop in three phases;
Short-Term (2012-2015), Mid- Term (2015-2020) and LongTerm (2020-2030). On the basis of population growth of
104
Figure 12: Sector wise split of land use Water

Requirement.
Proceedings
DCK, water demand for these three phases has been estimated
as shown in Table 4 and Figure 16 is represented its spatial
distribution.
By considering development plans it is estimated that for
shortterm development approximately 11.1 MG water will
be required that will serve the complete demand of the shortterm planned area. However, its water source development
will be tricky in the beginning plan; just for the short term
development plan average annual water resource development
requirement is around 3.7MG per year. However the actual
use of water will be started with the occupancy and use of land
for the respective purpose. It is estimated that in initial stage
only water is required for construction purpose and only 10 to
15 % of water will be required in DCK Camp site till the end
of short term development plan. It is better to develop around
1.5 to 3 MG water resource by the end of short term
development plan.
B.
Reduced water demand
A demand reduction measure serves to reduce water demand,
water losses, peak water demand, and non-essential water
uses. Reducing water consumption in the home, market and at
public places is a simple and easy way to decrease water and
energy bills and reduce households impact on the
environment. Conserving scarce water resources helps reduce
the need to dam rivers, reduce waste water produced and
treated at sewage plants, lower energy requirements for
treating and transporting water and waste water, and reduce
greenhouse emission.
1)
Strategy to Reduction water demand
For the efficient use of water at usage point is more effective

to save water and reduce overall municipal water demand.
There are many methods and strategy for water reduction, but
most feasible and easy to deploy are; Use of Water efficient
plumbing devices or fixtures and Recycling or re-use of grey
water
a) Use of Water efficient plumbing devices or fixtures
In the last two decades the use of water efficient plumbing
Figure 16: Water Demand for DCK development phases

fixture has been introduced throughout the world considering
water scarcity. It has been observed that water efficient
appliances in the home can reduce indoor water consumption
by 35-50%[31]. Typically used such devices and reported
water efficiencies are given in the Table 5.
Error! Reference source not found. gives figure for water
demand expected saving in water demand at DCK pertaining
to use of water efficient devices. It is clear from the above
table that the use of the aforementioned devices can save 44 %
water and the average water requirement will be reduced to 30
gpcd. The use of such devices can either be regulated in
building codes for their increased use. It is also suggested that
new home building should be guided to use these devices
through media campaigns. The land use like amenities and
commercial building give more saving by using water
conservation devices.
b) Recycling and re-use of Grey water
For grey water recycling, the overriding principle is safe usage
of recycling water. Therefore, we have not taken grey water
recycling where its safe usage can be questioned or its
consumption is not required. For plot sizes smaller than
500sqyrds, in any land use, have not been considered because
they do not have any green area and grey water usage for
landscaping is suggested only for large size plots. Health
facilities have not been incorporated for calculation of grey
water because of the hazardous nature of medical waste.
Therefore, grey water volume has been computed for
residential plots R5, R10 and R20, commercial plots (C5, C10,
and C20), mixed use (M5, M10, and M20), as well as planned
amenities. For this purpose, grey water shall be treated sites
and reused for landscaping. Amenities especially educational
institutes, produce a lot of grey water. This phenomenon
Table 5: Water demand using water efficiency devices

Name of efficient device
Frequency
use(per person)
Low-flush Toilet (1.6gpf)
Daily Water use with

water saving device
(gal/person)
5.1 flushes /day
Daily
Water
use
without
water
conservation device
(gal/person)
204
Low- volume showerhead

(2.5gpm)
5.3 minutes/day
Lowvolume
(1.5gpm)
Faucet
washing
Front-loading
machine (27gpl)
Water efficient dishwasher

(7gpl)
Total
of
Annual Water saving

(gal/person)
8.2
Daily Water saving

with water saving
device
(gal/person)
12.2
15.9
13.3
2.6
949
4 minutes/day
12
2190
0.37 loads/day
18.9
10
8.9
3249
0.1 loads/day
1.1
0.7
0.4
146
68.3
38.2
30.1
10987
4453
*Assumes conventional toilet at 4gpf, showerheads at 3gpm,faucet at 3gpm, washing machine at 51gpl, and dishwasher at 11gpl[1]
105
coupled with the fact that these amenities are usually

surrounded by many green spaced requires for the efficient
storage and usage of this grey water. Hence, it is suggested
that firstly smart taps and such fixtures be attached to lower
the water consumption and secondly to utilize the volume of
grey water produced to irrigate neighboring green areas.
c)
Proceedings
plot area 500sq yards or more. This is only possible when an

owner of houses installs separate drains for grey and non grey
water. This scenario will impact the reduction up to the 7.5%
of water required for the landscaping etc.
In adding to the first scenario of using grey water, if DCK
enforces the use of water efficient fixtures having rate of
water, saving up to 44% in the kitchen sink, faucets, showers,
flush, taps and laundry, with monitoring and auditing this
enforcement may produce a large impact on water demand
reduction (about 43.3%). In the third scenario only water
efficient devices are used that provide a moderate reduction in
water demand approximately 39.3%, which is slightly less
than second scenario. The results of these three scenarios for
different land use of DCK are shown in Table 6.
Water conservation at power plant:
Power house usually consumes a very large amount of high

quality water. The expected water requirement without water
conservation is about 79% of total utility water demand. This
high water usage is normally not provided through municipal
network. Therefore, according to the standard procedure, the
energy generation plant will manage its own water resource
based on deep well ground water, reclaimed water and limited
volume of municipal resources. It is expected that power plant
will not have designated quota from the municipal water
supply.
For water conservation; it will be recycled inside the plan and
it is further suggested that most economical water use practice
are deployed. The initial assessment is based on wet cooling,
which amount to 4.2 mgpd, whereas dry cooling process will
take only 0.58 mgpd. The other possibility is to recycle with
adopting wet cooling, which will give saving of 1.1 mgpd.
VI.
C.
Optimized Water Demand
The estimated water demand for DCK is 45MGD which can
be reduced by using recycled Grey water and use of efficient
devices. For this purpose water demand is calculated on the
bases of three scenarios. The first scenario is based on the grey
water use for landscaping and gardening within and outside
plots by the user of the building. For making it more practical
it is assumed to construct the grey water tanks separately for
storage and making it mandatory for all the residential,
commercial, mixed use and the amenities plot holder having
DISCUSSION AND CONCLUSION
Water scarcity has become a major problem in arid and semiarid regions therefore it is necessary to get rid this problem.
As implementation of the new water resource has least
probability due to its effect on the environment, therefore
alternate supply and demand reduction is the only solutions to
overcome water scarcity. Sustainable water resources
management is considered as an important application of GIS
and remote sensing, especially for the newly planned city.
Using this technology, this research not only help to estimate
the water demand for the new planned city with respect to its
land use, but also provided an optimized water demand by
using alternate water resources such as reuse or grey water and
using efficient devices.
The study results show that residential Land use has the
maximum water demand of 41% of total municipal water
demand, whereas the recreation and agriculture areas are
required comparatively the lowest water demand with only
1%. As DCK is considered as a first sustainable city of
Pakistan and for maintaining its high standard of life style
Table 6: Water Reduction by using three scenarios

Water Demand Reduction (%)
S- No
Land use Types
Scenario-1
Scenario-2
Scenario-3
(Use of Grey Water)
(Use of both grey water

and efficient devices)
(Use of efficient water saving

devices)
Estimated Water
demand (%)
Residential
41.3
10.2
49.5
44.6
Commercial
16.9
5.3
38.2
34.2
Mixed use(Res. Cum

Com)
16.4
5.4
44.6
41.9
Amenities
8.2
18.9
45.9
32.4
Recreation
1.6
0.0
85.7
85.7
Utilities
12.0
0.0
5.6
5.6
Transportation
2.2
00.0
00.0
0.0
Agricultural
Reserved Spaces
1.3
00.0
00.0
0.0
100.0
7.6
43.3
39.3
Total
and
106
comparatively high water demand is essential. Karachi water

and sewerage supply is the only water resource for DCK
which already supply insufficient water for Karachi city.
Reuse or recycling of waste water and use of efficient devices
serve as efficient and valuable ways to cope with the scarcity
of water and can reduce the water demand. The estimated
waste water for the DCK is 66% of the total water demand,
taken from the land use; residential, commercial, mixed use
and amenities, can reuse after the treated for landscaping and
agriculture outside the houses. For implementing these
alternate water resources, well-structured planning is
necessary. As for reuse of wastewater, black water drain
system should be separated with grey water and forcefully
implementation of water efficient fixture is essential for the
householder. There will be 45MGD water required for the
whole planning of DCK and water demands by using these
alternate resources are estimated. The results of the first
scenario based on reuse of grey water able of only 7.9 %
reduction of water demand. Whereas in the third scenario,
efficient device use show that efficient fixtures can reduce
water demand up to 39%. The second scenario, based on reuse
of grey water with the use of efficient devices, is considered as
an optimized water demand because of the maximum
reduction in water demand about 43%. After successful
deployment of the suggestion without failure to operate and
maintain the system could result in significant reduction in
water demand. Considering the scarcity of water in the area,
we expect that these critical management practices will be
adopted and implemented.
For further study this research can be extended to estimate the
portable and non-portable water use from the reused waste
water. This research provides a sustainable water resource
management system which is the requirement of the better life
style free from water scarcity which is expected to increase in
future.
VII.
ACKNOWLEDGEMENT
[4]
R. H. a. H. R. Vassilios A. Tsihrintzis, Fuentes., "Use of

Geographic Information Systems (GIS) in Water
Resources: A Review," Water Resources Management, vol.
10, pp. 251-277, 1996.
[6]
Z. X. Xu, K. Ito, G. A. Schultz, and a. J. Y. Li, "Integrated

hydrologica modeling and GIS in water resources
management.," Journal of computing in civil engineering,
vol. 15, pp. 217-223, 2001.
K. K. K. Mark R. Leipnik, and and Hugo A. Loaiciga,
"Implementation of GIS for Water resources planning and
management," Journal of Water Resources Planning and
Management, vol. 119, pp. 184-205, 1993.
A. Gemitzi, V. A. Tsihrintzis, O. Christou, and C. Petalas,
"Use of GIS in siting stabilization pond facilities for
domestic wastewater treatment," J Environ Manage, vol.
82, pp. 155-66, Jan 2007.
K. V. Kalivas DP, Karantounias G, " A GIS for the
assessment of the spatio-temporal changes of
[7]
[8]
[9]
the Kotychi lagoon, western Peloponnese, Greece.," Water Resourece

[10]
G. P. Panagopoulos, G. D. Bathrellos, H. D. Skilodimou,
and F. A. Martsouka, "Mapping Urban Water Demands
Using Multi-Criteria Analysis and GIS," Water Resources
[11]
O. Udovyk, "GIS for intergrated water resourses
management," Integrated Urban Water Resources
Management, pp. 35-42, 2006.
[12]
X. C. Daene C. Mc Kinney, "Linking GIS and water
resources management models:an object-oriented method,"
Environmental Modelling & Software, vol. 17, pp. 413425, 2002.
[14]
[2]
S. a. S. Ahmad, S. , "Spatial System Dynamics: New

Approach for Simulation of Water Resources Systems,"
Journal of Computing in Civil Engineering, vol. 18, pp.
331340, 2004.
L. Allen, J. Christian-Smith, and M. Palaniappan,
"Overview of Greywater Reuse: The Potential ofGreywater
Systems to Aid Sustainable WaterManagement," Pacific
Institute 654 13th Street, Preservation Park Oakland,
California.2010.
REFERENCES
[15]
[1]
F. A. Ward, "Decision support for water policy: a review of

economic concepts and tools," Water Policy, vol. 9, p. 1,
2007.
O. Udovyk, "GIS for intergrated water resourses
management," Integrated Urban Water Resources
ManagementNATO Security through Science Series., pp.
35-42, 2006.
[5]
[13]
The authors are grateful to DHA City Karachi, Pakistan and

Osmani & Company Pvt. Ltd. for sharing planning
information and other relevant literature related to the DCK.
Authors are also thankful to the GIS and planning team of
Osmani & Company Pvt. Ltd. for their great support.
VIII.
[3]
Proceedings
OCL, "Interim Report -1 , " Feasibility study for alternate

water supply stretegy including rechagre, reuse and small
dams " Osmani Company (PVT) limited and Pakistan
Defences Officers Housing Authority Karachi 2013.
F. Arbu s, . a. . Garc a- ali as, and . art ne Espieira, "Estimation of residential water demand: a stateof-the-art review," The Journal of Socio-Economics, vol.
32, pp. 81-102, 2003.
[16]
[17]
107
U. M. Shamsi, "GIS, Remote Sensing, and Master Water

Plan: A Case Study," In Computing in Civil Engineering
and Geographic Information Systems Symposium, pp. 695702, 1992.
P. Kirshen, Nvule, D., and Corliss, J, "Estimating urban
and suburban sewerage flows withassistance of GIS
technology," in Proc. ASCE Water Resow Plan. Manage.
Conf, Baltimore, MD, 2-6 August,, 1992, pp. 208-212.
P. R. a. B. S. Giguere, "Application of GIS to waste water
system mmmaster planning in Albuquerque," in ASCE
[18]
[19]
[20]
[21]
[22]
[23]
National Conference Civil Eng Appl. Remote Sensing GZS,

Washington DC,, 1991, pp. 115-119.
H. P. a. B. outal, D. , "Updating New York citys sewer
and water main distributionsystems: practical applications
of GIS," in ASCE National Conference Civil Eng Appl.
Remote Sensing GZS, Washington DC,, 1991, pp. 155-164.
OCL, "DHA city Karachi planning report," Osmani
Company (PVT) Limited and Pakistan Defences Officers
Housing Authority Karachi2012.
OCL, "DHA city Karachi ecology Report," Osmani
Company (PVT) Limited and Pakistan Defences Officers
Housing Authority Karachi2012.
U. Nation, "The Millennium Development Goals Report,"
UNITED NATIONS2012.
P. R. Philippe Loubet , Eleonore Loiseau , and V. BellonMaurel, "Life cycle assessments of urban water systems A
comparative analysis of selected peer-reviewed literature,"
Water Reseach, vol. 64, pp. 187-202, 2014.
T. Lu, "Research of domestic water consumption: a field
study in
Harbin, China," Master of Science of Loughborough University,

Water, Engineering and Development Centre Department
of Civil and BuildingEngineering, 2007.
[24]
[25]
E. M. M. Syed R. Qasim, Guang Zhu, Water Works

Engineering: Planning, Design, and Operation: PrenticeHall Of India Pvt. Limited, 2000.
M. Ahmad, "Pakistan Water and Sanitation Operators
Directory," Water and Sanitation program. 2012.
[26]
J. P. H. Miguel A. Morales, Kenneth R. Friedman, and

Jacoueline M. Martin., "Estimating commercial,
industrial,and institutional water use on the basis of heated
building area," Journal American Water Works Association
vol. 103, 2011.
[27]
[28]
R. Tabassum, "Title," unpublished|.

K. H. V. Durga Rao, "Multi-criteria spatial decision
analysis for forecasting urban water requirements: a case
study of Dehradun city, India," Landscape and Urban
Planning, vol. 71, pp. 163-174, 2005.
M. A. Al-Noaimi, "Decelopment of water resources in
Bahrain: A combined approch of supply demand analysis,"
PDh in School of Earth, Ocean, and Environmental
Sciences, University of Plymouth.
C. Qi and N. B. Chang, "System dynamics modeling for
municipal water demand estimation in an urban region
under uncertain economic impacts," J Environ Manage,
vol. 92, pp. 1628-41, Jun 2011.
D. Inman and P. Jeffrey, "A review of residential water
conservation tool performance and influences on
implementation effectiveness," Urban Water Journal, vol.
3, pp. 127-143, 2006.
[29]
[30]
[31]
108
Proceedings
Proceedings
Optimized Threshold Calculation based on received

signal characteristics for Blanking Non-linearity at
OFDM Receivers
Ferheen Ayaz
Zehra Ali
SUPARCO Institute of Technical Training, HR Dte Gen,

Space and Upper Atmosphere Research Commission,
Karachi, Pakistan
ferheen@msn.com
Space Electronics Wing,

Karachi, Pakistan
Zehra.aly82@gmail.com
AbstractIn satellite communications, OFDM (Orthogonal

Frequency Division Multiplexing) is evolving quickly as a wellestablished scheme. OFDM based communication system
performance, despite of its robustness and high efficiency, is
greatly affected by the addition of Impulsive Noise (IN). Additive
impulsive noise deteriorates the transmitted signal and overall
performance results poor when signal to noise ratio (SNR) and
symbol error rate (SER) are considered. Blanking nonlinearity
scheme is among the simplest methods to reduce the adverse
effect of impulsive noise at OFDM receivers. The performance of
blanking non-linearity is based upon threshold selection. In
blanking, the received samples whose magnitude exceeds a
certain fixed threshold are considered to be IN-affected and are
therefore set to zero. The value of threshold plays a very
important role in detecting the samples affected by IN. Setting a
fixed threshold does not always determine the IN-affected
samples truly as the OFDM distribution characteristics vary
throughout the signal. Several theoretical methods to optimize
the threshold level to improve system performance have been
reported. This paper reviews some of the threshold optimization
methods and proposes a new method based on some
characteristic distribution of the received signal including
median and peak. This proposed optimized threshold can be used
practically and results in higher or comparable Signal to
Impulsive Noise ratio than the other methods which are
incapable to be adopted in practical OFDM receivers.
KeywordsOFDM; impulsive noise; blanking non-linearity;
threshold; median; peak
I. INTRODUCTION
Orthogonal Frequency Division Multiplexing (OFDM) is a
rapidly growing modulation scheme and is widely adopted in
many wireless applications [1]. Performance of OFDM
modulation scheme for GNSS (Global Navigation Satellite
Systems) is one of the latest research topics [2]. Other
applications of OFDM include 4G mobile satellite systems [3]
and European Digital Terrestrial Video Broadcasting (DVB-T)
[4]. Despite of many advantages of OFDM communication
systems, the performance degradation in them still persists due
to high-power additive Impulsive Noise (IN) pulses and their
random occurrence in a channel [5]. DVB-T systems are
109
affected significantly by Impulsive Noise [6]. Performance

degradation in OFDM systems due to impulsive noise result in
high Symbol Error Rate (SER) and low Signal to Impulsive
Noise Ratio (SINR) [7].
A number of schemes to eliminate or reduce impulsive
noise from the received signal have been reported. These
schemes aim to increase the Signal to Impulsive Noise Ratio
(SINR) of the overall system. The simple approach to mitigate
impulsive noise from a received signal is to introduce a
memoryless non-linearity at OFDM receiver before
demodulation. Blanking, clipping and hybrid clippingblanking schemes have been devised using this approach.
Among all the non-linearity schemes, blanking is the easiest to
implement as it only converts the high amplitude samples
corrupted by IN to zero [7-8]. The blanking non-linearity
scheme revolves around the decision of whether the received
sample is affected by impulsive noise or not and then
changing it to zero. The decision is taken by a fixed parameter
known as threshold level or simply threshold. The threshold
level indicates if IN has corrupted the received signal sample
or the original sample value is retained at the receiver exactly
as transmitted. If the received samples amplitude exceeds the
threshold level, it is considered to be corrupted by IN and
therefore it is multiplied by zero. If the threshold level is
higher than the samples amplitude, it is assumed to be the
original sample value unaffected by IN and hence multiplied
by one to provide no change to the sample [9].
Fixing a value of threshold is very important as it defines
the overall SINR of the system. A low value of threshold may
blank the original signal samples which were initially not
corrupted by IN and a high threshold value may pass the IN
affected samples without blanking [10]. Fig. 1 illustrates the
two values of threshold to be used with blanking nonlinearity
on a signal affected by impulsive noise.
Since the performance of blanking non-linearity is greatly
affected by the selection of threshold, the choice of an
optimized threshold is a crucial decision. Optimized threshold
has been calculated in literature by different methods [11-12].
However, theoretical calculations of optimized threshold
cannot be used at practical OFDM receivers where probability

of IN occurrence is unknown. In this article, we review the
two threshold optimization techniques for blanking
nonlinearity; first is adaptive threshold optimization method
which maximizes the Signal to Impulsive Noise Ratio (SINR)
[11] and second is dynamic peak based threshold estimation
method [12]. We also propose a new method for calculating
threshold based upon median and peak of the received signal.
The proposed technique of threshold calculation has the
advantage over the other two schemes as it can be used at
practical OFDM receivers without the information of original
signal transmitted, probability of impulsive noise occurrence
and maximum SINR achievable.
12
Threshold 2
(Too high)
Amplitude
10
Threshold 1
(Too low)
2 .k.t
1 N 1
, 0 t T
Sk exp j
S
T
N k 0
S
(1)
Where Sk denotes frequency-domain signal, N is number

of subcarriers, j 1 , and Ts is active symbol interval.
Impulsive Noise (IN) is represented as ik and is included to the
transmitted signal using Bernoulli-Gaussian random process
as follows:
0 k N-1
(2)
with variance i2 (1/ 2) E[ gk ]. Hence, the signal-to-IN

2
ratio (SIR) can be given by; SIR 10log10 (1/ i2 ) . It is

worthwhile mentioning here that in this paper we represent
theoretical signal to impulsive noise ratio as SIR and practical
signal to impulsive noise ratio as SINR. The time-domain
signal ( rk ) received after addition of IN and AWGN can be
represented as
s(t )
Where bk 1 or bk 0 according to Bernoulli process

of sequence. The probability for bk 1 is p and bk 0
1 p . Whereas, g k is additive white Gaussian noise (AWGN)
0
0
In OFDM transmission, the time-domain form of

transmitted signal s(t) is found by applying inverse Fourier
transform to the frequency-domain signal by using
ik bk gk ,
OFDM symbols
Impulsive Noise
8
6
Proceedings
10
OFDM symbols corrupted by IN
12
4
x 10
Fig. 1. Two threshold levels considered for blanking IN-affected samples.
The organization of the remaining paper is as follows:

OFDM transmission system with blanking non-linearity
scheme is described in Section II. Section III overviews the
two threshold optimization techniques and proposes Median
Based Threshold (MBT). Section IV includes simulation
results in which MBT is compared with the other threshold
optimizing techniques with respect to Signal to Impulsive
Noise Ratio (SINR) and Symbol Error Rate (SER).
Conclusion is presented in Section V.
II. THE OFDM TRANSMISSION SYSTEM
if bk 0
s
rk k
,
sk ik if bk 1
The received signal is passed to the blanking non-linearity

block. The output of blanking non-linearity yk is then fed to
the OFDM demodulator for the further processing and is
represented as
yk rk dk ,
0 k N-1
(4)
Where d k represents the blanking non-linearity scheme and

is given as
1 r T
dk k
0 rk T
OFDM transmission is represented in the form of block

diagram as shown in Fig. 2.
(3)
(5)
Where T is the blanking threshold level and

k 0, 1, ..., N 1 . Optimizing the value of T for increasing
effectiveness of blanking non-linearity is described in the next
section.
III. THRESHOLD OPTIMIZATION
A. Overview of Threshold Optimization Techniques
An optimized threshold is very important for efficient
blanking of Impulsive Noise. Two threshold optimizing
methods are discussed in this section.
Fig. 2. OFDM transmission system with blanking scheme at receiver.
110
Proceedings
An Adaptive Threshold (AT) method is reported in [11]. AT

is the threshold which gives maximum signal to impulsive
noise ratio (SINR) according to (6).
AT max(SINR(T ))
(6)
AT for an OFDM signal is found by applying Blanking

nonlinearity to IN affected received symbols at all thresholds
varying from 0 to 15 with an increment of 0.1 and calculating
SINR each time. The threshold which gives maximum SINR
is denoted by AT as shown in Fig. 3.
20
AT at max(SINR)
SINR
15
10
Figure 4. Method to calculate Median Based Threshold.
MBT is given as:
0
0
10
MBT
15
Threshold
Another optimized threshold is known as Dynamic Peak

Based Threshold (DPBT) [12]. In DPBT estimation method
the peak of the OFDM signal is calculated before it is
transmitted through a noisy channel. At OFDM receiver,
blanking nonlinearity scheme is used with threshold equal to
the peak of the original signal as shown in equation (7).
0 k N-1
,0 k N-1
(8)
Where rk is received signal, max(rk) denotes the peak of the

received signal and is a threshold slope to be fixed at the
receiver. is set to achieve high SINR. Its optimum value
lies in the range for 4 to 6. In our simulations, it is kept 4
which was found to be optimized by using hit and trial
method. Setting an optimal value of is a better approach as
compared to selecting a fixed optimal threshold because a
threshold value may lie from 1 to 15 but the range of is
comparatively lesser and does not vary with different
probabilities of IN occurrence, unlike a fixed threshold.
Performance comparison of the three methods is shown in
the next section.
Fig. 3. SINR after applying blanking nonlinearity with respect to threshold

and AT, SIR=-15 dB, p=0.01.
DPBT max(sk (dB)),
max(rk ) median(rk )
(7)
Where DPBT denotes the threshold in Dynamic Peak

based Threshold Estimation Method.
Both of the optimized thresholds; AT and DPBT cannot be
used practically as it is impossible to find the exact peak of
transmitted signal and maximum SINR at OFDM receiver
[12].
IV. SIMULATION RESULTS

The OFDM system is simulated using MATLAB with 16QAM modulation, SIR=-15dB, number of subcarriers, N =
10^5, and i2 (1/ 2) E[ gk ] . Probability (p) of presence of
2
B. The Proposed Median Based Threshold (MBT)

Calculation Method
IN is taken as p=0.01, which suggests that 1% of the received

samples are corrupted by Impulsive Noise. The output SINR is
obtained by
Due to practical shortcomings of the techniques discussed

above, a new technique is therefore devised to calculate an
optimized threshold, known as MBT, by using the distribution
characteristics of the received signal which are its peak and
median. Fig. 4 illustrates the method to calculate MBT.
E[ sk 2 ]
SINR
E[ y s 2 ]
k
k
(9)
Fig. 5 shows the SINR achieved by three optimized
111
SINR(Gain / Loss)DPBT (SINRMBT / SINRDPBT )
(10)
SINR(Gain / Loss) AT (SINRMBT / SINRAT )
(11)
or equal to that resulted by AT .

1.2
MBT with AT
MBT with DPBT
1
0.8
SINR Gain/Loss (dB)
thresholds; AT, DPBT and MBT. It is noted that MBT mostly

results in a higher SINR than DPBT but in a lower SINR than
AT. However, the difference in SINR between MBT and AT is
not less than -0.64dB, which is quite negligible. This
difference in terms of SINR (Gain/Loss) is further illustrated
in Fig. 6. MBT results in a maximum of SINR gain of 1.2 dB
when compared with DPBT and 0.02 dB gain when compared
with AT . The relative SINR gain/loss is evaluated as
Proceedings
0.6
0.4
0.2
0
-0.2
-0.4
-0.6
-0.8
10
20
30
40
50
60
70
80
90
100
Number of simulations
Fig. 6. SINR Gain/Loss of MBT relative to DPBT and AT after applying
blanking nonlinearity.
9.5
Fig. 5. SINR from 100 simulations after applying blanking nonlinearity using
DPBT, AT and MBT.
From Fig. 5 and Fig. 6, it is clear that the performance of AT

is better than MBT and DPBT. Therefore, MBT is desired to
be closer to AT in order to achieve optimum performance. Fig.
7 shows that, as desired, the threshold values obtained by
MBT are near to AT which maximize SINR as compared to
the values of DPBT or peak of transmitted OFDM signal.
System performance of blanking nonlinearity with the three
optimized thresholds is also compared in terms of SER by
using
1
3
SER 1 1 2(1
)Q(
)
m 1
m
(12)
Where m is modulation index, Q denotes Q function and

denotes SINR.
Fig. 8 shows that MBT outperforms DPBT in resulting
SER of the system but the SER of MBT is slightly higher than
112
Threshold
9
8.5
8
AT
DPBT
MBT
7.5
7
6.5
16
16.5
17
17.5
18
18.5
SINR (dB)
Fig. 7. Thresholds (AT, DPBT, and MBT) relative to SINR after applying
blanking nonlinearity.
To check the authenticity of the proposed technique for

threshold optimization, SER was also obtained using AT,
DPBT and MBT with blanking non-linearity and making the
OFDM signal corrupted by impulsive noise with probability of
0.03. p=0.03 implies that 3% of the OFDM signal samples
would be affected by Impulsive Noise. The plot of SER with
respect to SINR in Fig. 9 shows that the performance of MBT
is close to AT. However, the performance of DPBT is
degraded, which means that DPBT is not feasible of all
probabilities of Impulsive Noise occurrence.
probabilities of Impulsive Noise occurrence. The proposed

method is sometimes incapable of achieving the maximum
SINR as that achieved by adaptive threshold optimization
method. However, the SINR achieved by this method is
always close to the maximum and acceptable in practical
scenarios.
-0.3
SER
10
10
Acknowledgment
-0.4
We gratefully acknowledge Training Division, HR Dte

Gen, Pakistan Space and Upper Atmosphere Research
Commission for their kind coordination in publishing the
article.
AT
DPBT
MBT
5.5
6.5
7
7.5
SINR (dB)
8.5
References
Fig. 8. SER with respect to SINR after applying blanking nonlinearity using
AT, DPBT and MBT, p=0.01.
SER
10
10
-2
-3
-4
16
16.5
17
17.5
SINR (dB)
18
S. H. Han, and J. H. Lee, An overview of peak-to-average power ratio

reduction techniques for multicarrier transmission, IEEE Personal
Communications, vol. 12, no. 2, April 2005, pp. 5665.
[2] X. Liu, M. Liang, Y. Morton, P. Closas, T. Zhang, and Z. Hong,
Performance evaluation of MSK and OFDM modulations for future
GNSS signals, GPS Solutions, vol. 18, no. 2, January 2014, pp. 163175.
[3] A. Vaneli-Corali, G. E. Corazza, G. K. Karagiannidist, P. T.
Mathiopoulosl, D. S. Michalopoulost, C. Mosquera, S. Papaharalabost,
and S. Sca1ise, Satellite Communications: Research Trends and Open
Issues, International Workshop on Satellite and Sapce Communication
(IWSSC07), IEEE, 2007, pp. 71-75.
[4] S.V. Zhidkov, Impulsive noise suppression in OFDM-based
communication systems, Consumer Electronics, IEEE Transactions,
2003, vol. 49, no. 4, pp. 944-948.
[5] M. Katayama, T. Yamazato, & H. Okada, A mathematical model of
noise in narrowband power line communication systems, IEEE Journal
on Selected Areas in Communications, vol. 24, no. 7, 2006, pp. 12671276.
[6] Y. Wu, Performance comparison of ATSC 8-VSB and DVB-T
COFDM transmission systems for digital television terrestrial
broadcasting, IEEE Transactions on Consumer Electronics, ,vol. 45, no.
3, 1999, pp. 916-924.
[7] S.V. Zhidkov, Analysis and comparison of several simple impulsive
noise mitigation schemes for OFDM receivers, IEEE Transactions on
Communications, vol. 56, no. 1, 2008, pp. 5-9.
[8] U. Epple, K. Shibli, and M. Schnell, Investigation of blanking
nonlinearity in OFDM systems, Iternational Conference on
Communications (ICC), June 2011, pp. 1-5.
[9] S.V. Zhidkov, Performance analysis and optimization of OFDM
receiver with blanking nonlinearity in impulsive noise environment,
IEEE Transactions on Vehicular Technology, ,vol. 55, no. 1, 2006, pp.
234-242.
[10] S.V. Zhidkov, On the analysis of OFDM receiver with blanking
nonlinearity in impulsive noise channels, Proceedings of International
Symposium on Intelligent Signal Processing and Communication
Systems (ISPACS 2004), November 2004, pp. 492-496.
[11] U. Epple & M. Schnell, Adaptive threshold optimization for a blanking
nonlinearity in OFDM receivers, Global Communications Conference
(GLOBECOM 2012), IEEE, December 2012, pp. 3661-3666.
[12] E. Alsusa & K. M. Rabie, Dynamic peak-based threshold estimation
method for mitigating impulsive noise in power-line communication
systems, IEEE Trans. on Power line, vol. 28, no. 4, October 2013, pp.
2201-2208.
[1]
AT
DPBT
MBT
10
Proceedings
18.5
Fig. 9. SER with respect to SINR after applying blanking nonlinearity using
AT, DPBT and MBT, p=0.03.
V. CONCLUSION
Two methods of optimizing thresholds of blanking
nonlinearity at OFDM receivers are reviewed. Both of the
methods are purely theoretical and infeasible to use in
practical OFDM receivers. One of the methods suggested peak
of transmitted OFDM signal as the optimized threshold but it
only worked for a single probability of impulsive noise
occurrence and did not give high SINR at other probabilities.
A new method of calculating optimized threshold based on
peak and median of the received signal samples is proposed
which can be used practically. This method results in higher
SINR and lower SER than the method which calculates
threshold by determining the original signal peak for different
This article is sponsored by Pakistan Space and Upper Atmosphere Research Commission
113
Proceedings
A Review on Mobile Wireless Communication Networks (0G to upcoming generations)

F.J.Sheikh 1
Dept. of CS & IT
University of Lahore, Lahore
farzanjavedsheikh@gmail.com
K.S.Mughal 1
Dept. of CS & IT
University of Lahore, Gujarat
khadija.shakeel10@gmail.com
I.T.Cheema1
Dept. of CS & IT
cheema.cheemas@gmail.com
AbstractAdvancements in mobile communication have

modernized the way people communicate, interact, and share
data and messages to each other. Latest technologies provide
these facilities to user in a short time. The technology of mobile
communication took its start from zero generation (0G) and now
fifth generation (5G) is near to overcome the world. Over as far
back as decade, remote innovation organization needs undergone
gigantic Growth. Zero generation is known as mobile radio
telephone, the predecessor of first generation. The 1st generation
(1G) has covered the essential portable voice. The 2nd generation
(2G) has announced the size and enabled the network to transfer
data, messages and MMS and followed by the third generation
(3G), which fulfill its mission to send data at high speed and
enough flexible to support 5 major radio technologies, this
further acknowledged by the fourth generation (4G) which
broadened the high velocity advancements and high versatility
through LTE (Long Term Evolution) and the fifth generation
(5G) mobile network correspond the framework with high data
transfer capacity and wide scope region. This paper will present
the evaluation of mobile wireless communication from 0G to
upcoming technologies. Evaluation include the wireless access
techniques used in all generations of mobile wireless, and
IPWireless technology used in all technology, purpose of LTE in
4G. What are the reasons and limitations that the new
generations or technologies are now replacing the old ones? This
paper also explores the need of upcoming generations that will
present the image of future wireless interaction in mobiles.
Keywords 1G; 2G; 3G; 4G; 5G; MMS; radio technologies;
LTE.
considered as a product for the improvement of extremely

consistent, tiny and solid state RF hardware.
The development of the mobile system, offer few benefits:
Function in least bandwidth, it also offer customer

and consumer fulfillment by the use of variety
adeptness feature.
In comparison of 'wired', wireless interaction are, in
many circumstances, low-cost to configure. Cost
feature is also reliable.
Because of the geographical area covered, number of
customers cannot build blocking probability. So
capacity structure must be high for mobile telephone.
II. ZERO GENERATION TECHNOLOGY (0G)
In earlier days, there were only one or two stations

available for mobile operators to set up a call. First modern
cellular mobile telephony technology was Mobile transistor
telephone system. Since these structures are staying the
prototype of the first generation of cellular telephone and
indicates as zero generation system (0G).
Technologies used in 0G methods involved Push to Talk
(PTT), Mobile Telephone System (MTS), Improved Mobile
Telephone Service (IMTS), Advanced Mobile Telephone
System (AMTS), Norwegian for Offending Land Mobile
Telephone (OLT), Public Land Mobile Telephony) and
Swedish abbreviation for Mobile telephony system(MTD).
I. INTRODUCTION
Data transmission without the use of improved electrical
conductor or wires over a distance is known as Wireless
communication. The area elaborate might be near (a limited
meter) or far away (thousands or millions of km for
broadcasting infrastructures). Some terms are often reduced to
Wireless due to indistinct framework that includes many
types of static, portable, movable, cellular phones, PDA and
wireless network. From the past few periods, technology of
mobile wireless (knowledge of many generas) named as 0G
to 4G. The operations of 5G technology or new upcoming
technologies can make improvement in usage of internet and
browsing data on WWW.
Respectively every genera has more or less principles,
capabilities, procedures and innovative functions over others
that separate them in others. The figure of cell phone
customers are growing with passage of time due to new
structures. At Bells test center, which was a cellular system
(1960-70), the development of wireless communication was
Mujahid Afzal2
Dept. of CS & IT
mujahid.afzal@uog.edu.pk
Push to Talk system are Half-duplex communication

that can be two way handshaking or present
connection that is called Push to talk and it is
working on personal computer to portable through
internet.
MTS framework was work bolstered in just as

instructions, if one calls by using public switch
telephone network (PSTN). Versatile administrator
who might request one's portable number for call and
the other numbers being called place the specific call.
When receiver was picked up from frame, IMTS

units form a specific dial bell and with this approach
it appeared as wired phone and less a mobile phone.
Almost 40-60 miles area was protected by IMTS in
diameter.
The frequency channel in bigger cities are large while

rural stations have one or two stations.
114
AMTS overawed all the problems that occur from

IMTS and works on 900 MHz
In Norwegh, OLT which worked on 160 MHz VHF

band was the first land mobile telephone network. In
regularity inflection it worked on 160 -162 MHz for
the mobile portion and in base station it is 168170MHz.
MTD include high speed internet structure with fixed

wireless facility that deals constantly on internet
access.
Proceedings
mobile phone program which use isolated frequencies in place

of individually discussion. It was changed commencing more
seasoned arrangements by "back-end" setup of calls usefulness
that permitted bigger telephones quantities to be bolstered
concluded geological region. In 1983, an aggregate of 40MHz
of range in the 800 MHz band was assigned by FCC for
AMPS. In 1989, additional 10 MHz has been charged to
AMPS because of ascent of cell framework limit. Uplink
frequency for AMPS remains 824- 849MHz and downlink
frequency stays at 869-894MHz. Independently AMPS
frequency is 30 KHz wide [16].
III. FIRST GENERATION (1G)

st
The 1 genera of mobile interaction was presented in the

1980 and analog transmission was used as communication.
The 1st mobile system (1979) developed by Nippon Telephone
and Telegraph (NTT) in Tokyo, Japan. Mobile Telephones
(NMT) and Total Access Communication Systems (TACS)
were two most popular system of analogue system [17].
Frequency division multiple access (FDMA) with channel
capacity of 30 KHz and frequency band 824-894 MHz use
incidence variation system for radio communication based is
derived from a technology called Advance Mobile Phone
Service (AMPS) [14, 15].
Fig 2. Architecture of Advance Mobile Phone Services (AMPS) [6]
IV. SECOND GENERATION (2G)

After first genera similarity mobile statement system 2G
mobile system was announced around 1991.concept of 2G
based on many base stations where each station circulated
consistently over the world to communicating with the users.
To converse more and more users various access procedures
are used i.e. FDMA, TDMA, and CDMA [10]. 2G technology
make procedures of compression decompression procedure
(codec) and family members of this generation are 2G (GSM,
GPRS, EDGE).
Digital signal usage among recipients and tower build the
framework ability in two forms
1) Digital sound data be able to smashed and multiplexed
proficiently in this way consenting more data is delivered into
the equal quantity of radio data transmission.
Fig 1. 1G Mobile Phone [6]
The initial step of mobile communication is NTM. It

is in the result of heavy and increased usage of physical
cellular communication. The variations of NMT are NMT-450
and NMT-900. The figures specifies usage of band rate
correspondingly. NMT-900 contain other Channels as
compared to NMT-450. NMT cells cast-off full duplex
program. Permitting in real-time acceptance and vocal sound
communication [14].
2) Less radio power is released while using digital

systems. Therefore creation of small size of and altering better
mobiles in similar quantity of space. Two standard was used
for 2nd genera technology CDMA and TDMA on center of
multiplexing. In drill, TDMA and CDMA systems share with
FDMA. TDMA expression is utilized to assign frameworks,
which partition the channel into frequency openings, later on
isolate every rate space into various time openings. Also,
CDMA is truly a cross breed of CDMA and FDMA where the
station is initially partitioned into recurrence openings.
AMPS was a reporter cellular telephone framework typical

creates in 1983 by Bell Labs, America. It was an initial genera
115
Proceedings
permitted 3G links in personal computers, digital gaming

equipment, tablets, and smart phones including all devices that
are able to take benefit by using high and fast quality of
internet usage. Working capacity of devices and functionality
of devices can be increased many times with the usage of fast
and reliable internet connection. Better safety and less threat is
offered while connecting and transferring data to other
wireless devices by using 3G [2, 3]. Three basic technologies
are covered by 3G that is CDMA2000, TD-SCDMA - Timedivision Synchronous Code-division Multiple Access and WCDMA (UMTS) Wideband Code Division Multiple Access.
The vital role of 3G knowhow is, it offers high transfer of data
transmission and improved speed. Packet-switching technique
is used in 3G that is considered more realistic and also much
faster as compared to the previous used technologies and
techniques, issue with it is it requires a totally different
groundwork as compared to the 2G systems.
GPRS is a cell phone wireless technology established

amongst prototype, 2G, and predecessor, 3G) [12]. Its a radio
innovation for GSM systems that upgrades bundle changing
over conventions, set up period for ISP systems, choice to pay
according to the sent data, instead of association period. In
packet transferring method, material to be delivered is
damaged into packets, size of the packets is in kb, after that
these are moved through network among various ends, each
data packet contain address of destination. Data transmission
and network connection is the provision of GPRS.
EDGE is an ordinal cell phone technique that actions as a
bolt-on enrichment to General Packet Radio Service (GPRS)
networks. It is ideal as compared to GSM because it carries
packet switch data and circuit switch data [11]. Black berry
N97 and N95 mobile phones are famous because of EDGE
technology. By using EDGE technology, no additional
hardware and software resources are required for making
EDGE technology work. Any of the extra expense are not
required for developing such technology [13].
Fig 5. CDMA Architecture [6]
Mobile communication market is increasing on a

very fast pace, usage of mobile phone industry is increasing in
leaps and bounds. Almost one billion customers globally are
benefiting from cellular phone industry, in which two-thirds
are GSM customers. One of the firmest growing wireless
technology is known as CDMA, it is developing its
subscribers base signify cantle around the world. Nowadays
CDMA subscribers are crossing the line of 200 million users.
3G networks enable network workers to deal customers an
extensive variety of latest progressive facilities and also
realizing better and superior connection ability through
healthier spectral efficiency.
Fig 4. GPRS Architecture [6]
V. THIRD GENERATION (3G)

3G is the innovative generations for cellular interaction
facilities, its services area centered on procedural values of
IMT-2000 containing the consistency, accuracy and also speed
(rate in which data is transferred) that is minimum 200kb per
second [11]. Outside of cell phone, more data transfer ratio
116
Proceedings
VI. FOURTH GENERATION (4G)

4G technology uses 3G as a inheritor, developed for the
broadband capabilities with the services of 3G to enable the
high data rate streaming i.e.20 Mb/s per customer, Quality of
services, Fast Speed, Faster Response Time, Routing
efficiency, Higher Bandwidth and Lower Power consumption.
Speed for low mobility is 1 gigabit per second (Gbit/s) and
100 megabits per second (Mbit/s) for peak speed. NTT Do Co
Mo was first successor to achieve. Downloading of packet
transmission at a moving speed of almost 20km/h is 1Gbps in
Tokyo, Japan on June 23rd, 2005 [2]. 4G technology is
combination of IP-Based voice, Seamless access,
Personalization, Quality of services. Its main purpose is to
access network anytime, anywhere and anyhow. 4G wireless
communication technology is furthermore discussed as
MAGIC. MAGIC word stands for mobile multimedia, anywhere, global mobility solutions over, integrated wireless,
customized services.
Need for this platform is to integrate pervious mobile
technologies entirely which are present that is GSM, GPRS,
IMT-2000,
Wi-Fi,
Bluetooth,
WCDMA, HSDPA, CDMA2000 and EVDO,
provide
correspondence to user expectations for various services.
Mobile WiMAX regularities and Long Term Evolution (LTE)
rules were deployed for 4G but failed to satisfy the IMTAdvanced requirement, may well though measured as 4G.
Later on LTE Advanced (latest release of LTE) standardized
by the 3GPP (Third Generation Partnership Project) and
802.16m uniformed by IEEE (that is WiMAX). Demand of
IMT-Advanced for the generations of interactive mobile
services by providing a global platform is increased to gain
unified messaging, quicker information admittance, improved
roaming competencies and broadband multimedia [16].
A. LTE Advanced
Fig 7. LTE Advance Architecture [17]
VII. FIFTH GENERATION (5G)

5G (Mobile and wireless Network) can make the perfect
wireless real world without any limitation is known as World
Wide Wireless Web (WWWW).
Beyond the 4G/IMT-Advanced standards 5G symbolizes the
next utmost chief level of cellular phone technology genera.
5G is still not considered and used as an official term on
behalf of any specific requirement and design and also not
publically used by telecommunication companies or
standardization such ITU-R (International Telecommunication
Union-Recommendation), 3GPP, WiMaxForum (Worldwide
Interoperability for Microwave Access) [2]. Every new release
of technology enhances the system performance with some
advance capabilities. 5G technology must be something else
than previous technologies, whose purpose is not only to
increase the throughput but also consider the other
requirements.
LTE is not such a new technology but some essential

enhancements in LTE to improve existing LTE network. It
was officially candidate of ITU-T (International
Telecommunication Union- Telecommunication), meeting the
requirements of the IMT-Advances standard and also
standardized by 3GPP. LTE Advanced has 30bps/Hz peak
spectrum downlink value and 15 Bps/Hz peak spectrum uplink
value, can be enhanced if multiple-input multipleoutput (MIMO) (that is antenna arrays) remain used. Both
coordinated multipoint (COMP) and MIMO remained
recognized such as the crucial procedures for LTE [2].
117
Improved coverage of geographical zone and high

rate for data transfer at the edge of the mobile phone
Truncated power and battery consumption
Accessibility of numerous data transmission tracks
Nearby 1 Gbps data rate is definitely thinkable
Additional Security
Energy effectiveness and spectral efficiency are good
[17]
Proceedings
IX. UPCOMING GENERATIONS

Upcoming technologies will provide the real image of
wireless world with less boundaries. Artificial intelligence and
ubiquitous network will provide universal computing from
anywhere, anytime, and any device. User can instantly linked
to several technologies and perfectly move between them.
X. COMPARISON
We compare all the technologies of Mobile Technology in
term of data bandwidth, standards, technology, multiplexing,
switching see table 1. [6]
XI. CONCLUSION
5G mobile technology will be powerful and

demanding in future due to including most powerful features.
Also it will change the concept of using mobile phones with
more bandwidth. Customers certainly not have experience of
such great and fast tracking technology. The architecture of
the fifth generation consists of cellular networks
interoperability (contains of a customer station and number of
autonomous radio access technologies (RAT)) and wireless IP
based model [5]. User terminal has a significant role in this 5G
technology system. Proposed 5G architecture is shown in fig
8. 5G technology deal with camera, large phone memory,
MP3 recording, dialing speed, video player, audio player and
many creative, innovative and imaginative technologies [13].
Wireless communication technologies has captured the

market since its first start. This study provides the better
understanding of the previous and present technologies along
with their performances, portals, advantage and disadvantages
of one generation to other and also predict incoming future
technologies.
.
REFERENCES
[1]
[2]
[3]
[4]
[5]
[6]
[7]
Fig 9. Functional Architecture for 5G- Network [18]
[8]
VIII. SIXTH GENERATION (6G)

The 6G technology will combine the 5G satellite network
and mobile wireless system. Satellite network include
telecommunication satellite, navigational satellite, imaging
satellite, environmental information system. These system
used for data, voice, video broadcasting and internet; weather;
environmental information Collection; and global positional
system (GPS) respectively [2]. At the same time Fifth
generation will focus on user terminals but it is necessary for
terminals to have access of different wireless technologies.
Roaming and handoff will be the big issue of 6G because of
the different network stations. Solving the problem of handoff
and roaming is still a question. But the charges of mobile call
will be relatively high [8].
[9]
[10]
[11]
[12]
[13]
[14]
118
F. R. Y. Chengchao Liang, "Wireless Virtulization for Next Generation

Mobile Cellular Networks," in IEEE Wireless Communications,
February 2015.
P. Sharma, "Evolution of Mobile Wireless Communication Networks1G to 5G as well as Future Prospective of Next Generation
Communication Network," in International Journal of Computer
Science and Mobile Computing, August 2013.
E. M. I. A. E. U. M. A. Engr. Muhammad Farooq, "Future Generations
of Mobile Communication Networks," in Academy of Contemporary
Research Journal, January 2013.
A. G. Roopali Sood, "Digital Society from 1G to 5G: A Comparative
Study," in International Journal of Application or Innovation in
Engineering & Management (IJAIEM), February 2014.
R. Rajan, R. Baweja and R. Kumar, "Review on Different Multiple
Access Technique Used in Wireless Communication," in International
Journal of Research (IJR), November 2014.
P. S.Venkata Krishna Kumar, "A Study of Wireless Mobile
Technology," in International Journal of Advanced Research in
Computer Science and Software Engineering, January 2014.
V. G. APANA BHAT, "EVOLUTION OF 4G: A STUDY,"
International Journal of Innovative Research in Computer Science &
Engineering (IJIRCSE), February 2015.
A. S. J. R. Sarmistha Mondal, "A Survey on Evolution of Wireless
Generations 0G to 7G," International Journal of Advance Research in
Science and Engineering- IJARSE,.
A. A. J. I. Nabeel ur Rehman, "3G Mobile Communication Networks,"
EXPLORE SUMMER 2006, 2006.
N. Schmitz, "The Path To 4G Will Take Many Turns," 1 March 2005.
[Online]. Available: http://electronicdesign.com/4g/path-4g-will-takemany-turns.
A. G. R. S. O. Z. Xichun Li, "The Future of Mobile Wireless
Communication Networks," 2009 International Conference on
Communication Software and Networks, 2009.
A. R. Mishra, Fundamentals of Cellular Network Planning and
Optimisation: 2G/2.5G/3G... Evolution to 4G, John Wiley & Sons
2004, 2004.
A. F. Molisch, Wireless Communications, Wiley, 2010.
F. ADACHI, "Wireless Past and Future--Evolving Mobile
Communications Systems," IEICE TRANSACTIONS on Fundamentals
of Electronics, Communications and Computer Sciences, 2001.
[15] D. Y. L. D. J. S. D. Amit Kumar, "Evolution of Mobile Wireless

Communication Networks: 1G to 4G," International Journal of
Electronics & Communication Technology, December 2010.
[16] A. Pashtan, "wireless terrestrial communication: cellular telephony,"
TELECOMMUNICATION SYSTEM AND TECHNOLOGIES, 2006.
[17] G. M. K. A. Arun Agarwal, "The 5th Generation Mobile Wireless
Networks- Key Concepts, Network Architecture and Challenges,"
Proceedings
American Journal of Electrical and Electronic Engineering,, vol. 3, no.

2, pp. 22-28, February 2015.
[18] [Online].Available:http://www.prismatelecomtesting.com/products/prod
ucts-overview/radio-interface-simulation/uesim/.
TABLE 1. COMPARISON OF MOBILE TECHNOLOGIES

Generation Features
1G
2G
3G
4G
5G
Years
1980s
1990s
2000s
2010s
2020s
Data Bandwidth
2kbps
64kbps
2Mbps
200Mbps
1Gbps
Standards
AMPS
TDMA, CDMA,GSM,
GPRS
WCDMA
Single unified
standards
Single unified standards
Technology
Analog cellular
Digital Cellular
Broadband with
CDMA, IP
Technology
Unified IP & Seamless

combination of
broadband,
LAN,WAN & WLAN
Unified IP & Seamless

combination of
broadband, LAN,WAN
& WLAN,WWWW
Service
Mobile technology
(voice)
Digital Voice, SMS,

Higher Capacity
Packetized
Integrated high
quality audio, video
& data
Dynamic information
access, wearable
Devices
Dynamic information
access, wearable
Devices with AI
capabilities
Multiplexing
FDMA
TDMA,CDMA
CDMA
CDMA
CDMA
Switching
Circuit
Circuit & Packet
Packet
All Packet
All Packet
119
Proceedings
Intelligent Detection of Distributed Flooding Attack

in Wireless Mesh Network
M.Altaf Khan
Institute of Information Technology
Kohat University of Science & Technology
Kohat, Pakistan
altaf_apc@yahoo.com
Amjad Mehmood
Kohat, Pakistan
Shafi Ullah Khan

Kohat, Pakistan
skkust@hotmail.co.uk
Zeeshan Iqbal
Kohat, Pakistan
Zeeshan.zn@gmail.com
amjadiitkust@gmail.com
AbstractMulti-hop wireless mesh network (WMN) is a

growing and promising technology that fulfills the requirements
of next generation wireless networking by providing an extensive
assortment of applications that were not assumed to be supported
by a wireless network. In WMN, a Gateway is the only way of
integrating it with other wireless and wired networks. Due to this
critical role, gateways in WMNs are targeted by diverse security
threats. A distributed denial of service (DDoS) attack is one of
the prime security threats to WMN. DDOS attack is an explicit
attempt to prevent the legitimate users from accessing their
service. In a distributed flooding attack, gateway can be easily
exploited by the attackers with the aim to reduce the available
network bandwidth for the valid users. Hence network
performance is reduced and also destruct the services provided to
the end users. Numbers of detection mechanisms have been
provided by researchers in this regard. However, due to
advancements in the attacking tools and techniques, more
intelligent efforts are required in terms of precision and
competence, to protect the WMNs from such attacks. In this
paper, an intelligent mechanism called Distributed Flooding
Attack Detector (DFAD) is proposed with the aim to detect
distributed flooding attacks in WMNs. Network simulator (NS2)
is being used to evaluate the performance of DFAD.
KeywordsWireless Mesh Network, Distributed Denial of

Service Attack, Mobile Ad-hoc Network, Artificial Neural
Network, Back-Propagation
I. INTRODUCTION
A multi-hop and multi-radio wireless mesh network
(WMN) is a decentralized, dynamically self-organized and
self-configured network. Connections among the nodes in
WMN establish without any central controlling body using
partial or complete mesh topology [1]. Two types of nodes are
existed in a typical WMN i.e. mesh node (MN) and mesh
router (MR). MN may either be static or mobile having one
wireless interface. MRs forms a multi-hop backbone of the
WMN by providing self-configured and self-healed links.
Sometimes, in WMN, a MR may have both wired and wireless
interfaces which enables it to provide functionality of gateway.
Unlike conventional routers, MR exhibits some additional
routing functions to bear mesh networking. Moreover, due to
multi-hop support, mesh routers provide more coverage with
low transmission power than conventional routers. WMNs can

be classified into three main architectures i.e. infrastructurebased, infrastructure-less and hybrid WMN [2]. Infrastructurebased WMN is composed of both MNs and MRs.
Infrastructure-less WMN seems to be a pure mobile ad-hoc
network (MANET), because it is consisted of only MNs that
may also carry out routing and self-configuration to form a
peer to peer wireless network. Hybrid WMN is composed of
MNs and MR, where MRs are used to form backbone of WMN
and MNs offer routing abilities in order to achieve an enhanced
coverage and connectivity in the interior WMN. In both
infrastructure-based and hybrid WMNs, through mesh gateway
(MG), a node in WMN can communicates with other existing
networks. Hence, MG plays a vital rule for instigation or
termination of traffic in WMN. So it is mostly targeted by
intruders during different attacks. Once MG is being
compromised, all the provided services of WMN are affected
and hence degrades its performance.
WMNs have made life of its users easier. However, some
serious security threats may be faced by the WMN that may
compromise the confidentiality, integrity or availability of a
WMN [3]. Confidentiality is being compromised through
several attacks like eavesdropping, traffic analysis, corrupt
access point, war driving attack, and brute force attack. Blackhole, worm-hole, grey-hole etc. are the attacks through which
integrity of WMNs are being compromised. Availability is
compromised by many DOS attacks including TCP SYNflooding attack, UDP-flooding attack and smarf attack, where
intruders try to prevent normal services or to make the
network resources unavailable for the legitimate users. When
a DOS attack is implemented by thousands of attackers
simultaneously, with the aim to increase the severity of attack,
it is identified as DDOS attack. An intruder launches such
attack by sending a massive bulk of ineffectual packets
concurrently towards victim, through huge number of
compromised nodes also known as zombies [4]. These
zombies may either be internal or external to the victims
network. Distributed flooding attack (DFA) is a one of the
classical approach to launch DDOS attack in WMNs using
120
tools like Trinoo, Tribe Flood Network (TFN), Shaft, Knight

and Trinity [5].
Anomaly based DDoS attack is being detected in [6].
Features of malicious packets are analyzed using radial base
function neural network (RBFNN). A vector of seven features
is being used for activation of RBFNN and classification of
incoming packets flows into classes of attack traffic or normal
flows. Proposed method is evaluated through datasets of
UCLAD. Proposed method classified the incoming traffic into
class of normal or attack traffic, but it cannot classify and
identify the type of attack. A statistical approach, based on
values of numerous features is implemented in [7] to detect
attack traffic. Proposed method used a module for features
extraction and then extracted features are stored in a database
for identification and classification of normal and malicious
packets flows. A novel approach is presented in [8] to detect
flooding and IP spoofing attack. Prevention of flooding attack
is accomplished through considering clock values of
individual node of network. A combined data mining
approach is used in [9] against DDoS attack for detection of
protocol anomaly. In the proposed method for extraction of
features network traffic is utilized and then a classification
algorithm based on data mining is used to cluster incoming
traffic into normal and attack flows.
For detection of such attack numerous techniques are
proposed so far, but the techniques that are based on
intelligent schemes are proved to be more proficient
alternatives in this regard [10,11,12] . Still there is a need to
propose more mechanisms to handle this attack. Our proposed
mechanism is more intelligent and reliable then the
mechanism proposed earlier for detection of DFA in WMNs.
Soma
Dendrites
Axon
Nucleus
Synaptic
Terminals
Fig.1. Structure of a biological neuron
Following the same concepts, ANN is a network of

interconnected nodes (neurons) that provide required solution
after performing parallel processing. Structure of earliest ANN
called perceptron, is given in the Fig.2.
X1
X2
Output (0/1)
X3
Xn
Fig.2. Structure of a perceptron
Using equations (1,2,3), If Sum of the product of binary

inputs (Z1, Z2. Zn) and their associated weights (W1,
W2....Wn) is found greater or equivalent to 0 then f i.e.
an activation function will turn out 1 as output if not, 0
will be the output of the perceptron [14].
(1)
II. PROPOSED MECHANISM

Due to self-configuration and dynamic changes in
topology of WMN, ANN is considered to be the most suitable
technique amongst all other intelligent techniques to detect
distributed flooding attack [5]. Our proposed mechanism i.e.
distributed flooding attack detector (DFAD) is based on ANN
technique and is adaptive in nature.
ANN used the concepts of biological neural networks
found in human brain. With the assistance of this network a
brain involves millions of interconnected neurons to fully
control the human body. Fig.1 depicts the basic structure of a
biological neuron. It illustrates some important components of
a neuron. Actual cell body of the neuron is constructed by
Soma. It is composed of a nucleus and plasma. Information
regarding to innate traits is provided by the nucleolus and
plasma generates material required by neurons. Dendrites and
Axons are used during communication with other neurons.
Through synaptic terminal an Axon spreads out received
signal to other neurons [13].
Proceedings
))
(2)
(3)
Two types of structures are existed in ANNs i.e. feedforward and feed-back neural networks. Main difference
between both structures is the flow of signals. Feed-forward
neural network allows information flow in just forward
direction while a feed-back neural network allows information
to be flowed in both forward and backward directions[15].
These structures are mainly distributed in three main layers:
input-layer (IL), hidden-layer (HL) and output-layers (OL). IL
receives input for the network and the OL provides the results
of the network. HL offers the neural network with its aptitude
to generalize. Moreover, HL may be composed of more than
one layer. There may be different number of neurons in each
layer. As quantity of layers and nodes in HL increases,
complexity of ANN will be increased. On the other hand,
probability of accurate learning is increased. A feed-forward
ANN is illustrated in Fig.3.
121
IL
HL
value of weights and biases. denotes momentum and

represents the learning rate. In equation (7) and (8),
and
represents the desired change in weights associated with
hidden-to-output layers or input-to-hidden layer connections
and biases respectively in the backward-pass.
In our DFAD, we used multiple layer perceptron (MLP)
having three inputs nodes in the IL, because we aimed to
check the MG after each five seconds. Each input value
represents received number of packets. After trying different
number of nodes and different number of layers in the HL,
finally we got our results with higher detection rates and
lower false rates, using one HL having four nodes in it. As
we are interested to distribute the incoming flows at MG into
three different categories therefore we used three nodes in
the OL to generate one of the desire output i.e. [1,0,0],
[0,1,0], [0,0,1] to represent normal, low distributed flooding
(LDFA) attack and sever distributed flooding attack (HDFA)
flows respectively.
During training of DFAD, a vector of eight (08) input
values is provided. Arrangement of the input vector is
[4,3,7,0,0,1], where [4,3,7] are the amount of packets
received in each second and the remaining values of the input
vector i.e. [0,0,1] represents the target output for the given
input values. We used three different target outputs with
each input vectors i.e. [0, 0, 1], [0, 1, 0] and [1, 0, 0], Where
[0, 0, 1] is imagined as normal flows, [0, 1, 0] is considered
to be an intermediate distributed flooding attack and [1, 0, 0]
is expected to be a high distributed flooding attack flows. A
minute detail of training dataset is given in the Table.1.
OL
Inputs
Fig.3.
Feed-forward artificial neural network
Learning of ANN can be performed through either

supervised or un-supervised learning algorithms. In
supervised learning, an output is provided with each input
during training while in unsupervised learning algorithms,
target outputs are not provided with input data during
training. Hence ANNs perform data compression or
clustering to categorize patterns of the same attributes into
same output clusters [16].
In feed-forward ANNs, most commonly used supervised
learning algorithms may include delta rule, perceptron and
back-propagation (BP). For learning a feed-forward ANN,
BP algorithm is most commonly used amongst entire
algorithms.
Delta rule and perceptron algorithms are useful to
resolve merely linear problems, while BP algorithm is able
to provide learning in complicated non-linear problems.
During training, in BP based ANN, signals are transferred in
forward direction i.e. from input to OL. Error between actual
and target output in the OL is calculated in each pass. In BP,
learning process includes forward-pass and backward-pass.
Input data is moved from IL to OL in forward-pass. Most
commonly sigmoid activation function is used in BP
algorithms to generate results from each HL and OL node by
using following equation (4). Output of this sigmoid
function lies between 0 and 1
( )
(4)
Error (E) between target output (T) and actual output (O)
is calculated at the end of each forward-pass. Gradient
descent technique is used to reduce this error E between Oi
with respect to Ti. The value of E, gradient descent and
required changes in the weights of each connection values
are computed through following equations respectively.
(5)
(6)
(
(7)
(8)
In equation (6), GrD represents value of gradient

descent, its value concludes that either to raise or reduce
Proceedings
Input
Vectors
4, 3, 7, 0, 0, 1, 31, 34, 36, 0, 1, 0, 42, 51, 65,

1, 0, 0, 4, 6, 12, 0, 0, 1, 33, 26, 22, 0, 1, 0,
46, 54, 57,1, 0, 0, 16, 6, 17 0, 0, 1, 37, 35,
46, 0, 1, 0, 54, 58, 56, 1, 0, 0,....
Weights
and
Biases
0.11, 0.21, 0.13, 0.14, 0.25, 0.36, 0.17, 0.8,

0.19,-3.0, -4.0, -2.0, 1.6, 1.8, 1.3, 1.6, 1.2,
1.1, 1.7, 2.1,2.3,-2.3, -7.0, -5.0
Table.1. Training dataset
Assortment of values for initial weights and biases is an

actually very tough job. Numbers of forward and backward
passes to generate the desired outputs from the network depend
on these values.
In training, maximum 100 epochs are used with each input
vector to minimize the error between target output and actual
output. Dataset used to train the network, is consisted of 5000
of input vectors, having all three types of flows, mentioned
above. After completion of training, updated values of weights
and biases are used by DFAD to distribute the input vectors in
desired category in testing phase.
Algorithm of proposed mechanism is given in Fig.4. and
Fig.5. Main difference between training and testing phase is
that, both passes i.e. forward and backward passes of BP
algorithms are used in training phase while in testing phase
only forward-pass is used. Moreover, in testing phase input
vector is entered in the system without target output values.
122
=====================================
Testing:
=====================================
1. Assign all Updated weights and biases
obtained from training phase
2. Input Current pattern to the network
3. // Propagated the input forward through the
network:
4. For node in the Hidden layer
i. Calculate the sum of product of
weights and inputs to the
node using Eq.2
ii. Add the biase of each node to the
calculated sum
iii. Calculate the output using Eq.4
for each node
5. Next
6. For node in the Output layer
weights and inputs to the
node using Eq.2
calculated sum
for each node
7. Next
8. // Output Layer: Node1s output: out1,
Node2s output: out2,Node3s output: out1;
9. IF(out1>out2&&out1> out3)
10.
printf (Normal flow);
11. elseif (out2> out1&&out2> out3)
12.
printf(LDFA-Traffic);
13. else
14.
printf(HDFA-Traffic);
15. endif
=======================
Training:
==================================
1. Initialize all weights and biases with small
random numbers
2. For Input vectors in the training set
3. Input Current pattern and target output to the
network
4. // Propagated the input forward through the
network:
5. For node in the Hidden layer
weights and inputs to the node using
Eq.2
calculated sum
iii. Calculate the output using Eq.4 for
each node
6. Next
7. For node in the Output layer
weights and inputs to the node using
Eq.2
calculated sum
foreach node
8. Next
9. Calculate sum of error between target and
actual output using Eq.5
10. IF ((maximum number of iterations (epochs)
< 100) && (Error>=0.25))
11. // Propagate the errors backward through the
network
12.
For node in the output layer
i. Calculate Gradient value for node in
the output layer using Eq.6
ii. Update each node's weight and biase
values in the output layer using Eq.7
and Eq.8
13.
Next
14.
For node in the hidden layer
i. Calculate the Gradient of node's in the
hidden layer using Eq.6
ii. Update each node's weight and biase
value in the hidden layer using Eq.7
and Eq.8
15.
Next
16.
With updated weights and biases repeat
from Step-5
17. endif
18. Next // select next input vector for training
Proceedings
Fig.5. Testing phase
Dataset used to train the network, is consisted of 5000 of

input vectors, having all three types of flows, mentioned
above. In training, maximum 100 epochs are used with each
input vector to minimize the error between target output and
actual output. After completion of training, updated values of
weights and biases are used by DFAD to distribute the input
vectors in desired category in testing phase.
III. SIMULATION RESULTS
NS2 [17] is considered to be a best choice for simulations of
wired and wireless networks therefore we carried out our
simulations in NS2.34. Effectiveness and accuracy of DFAD
is tested by performing simulations on NS2. Following
parameters are set for our simulations.
Fig.4. Training phase
123
Simulation time
60 sec
Number of nodes
100
MG
01
MR (Fixed nodes)
09
Mobile nodes
90
Speed of mobile nodes
5 m/sec
Simulation area
500 X 500
Nodes transmission range
250 m
Proceedings
Table.2. Simulation parameters
To perform training of the DFAD, 40 nodes are used to

generate flows with different data rates towards gateway. To
generate normal packets flows, we set transmission rates of
these nodes to 8Kbps, low and high distributed flooding attack
flows is generated with the transmission rates of 15 Kbps to
35 Kbps respectively. Network animator (NAM) i.e. an
animation tool usually used by NS2 to generate a view of the
simulation is being shown in Fig.5. It shows that packets flows
garneted by both mesh nodes and attacking nodes will reach to
the gateway through mesh routers. The nodes which are not in
the range of mesh routers will use nearby mesh nodes to
transfer the packets towards the gateway.
Fig.6. Output of NAM
After each experiment, DFAD distributed the received flows

of mesh gateway accordingly as Normal, LDFA or HDFA
attack flows. Distribution of both TCP and UDP flows in each
experiment is given in Fig. (7, 8).
Packets
Normal- Flows
LDFA-Flows
HDFA-Flows
Time (Sec)
Extensive simulations with different data rates were

carried out for both UDP and TCP flows. Simulation time is
60 seconds for each experiment. Initial experiment is
performed with 40 nodes where each node transmitted data
with rate of 08Kbps.In second experiment, there were 30
normal nodes and 10 malicious nodes (attacker). Each normal
node transmitted with the rate of 08 Kbps and malicious nodes
transmitted with 20 Kbps. In our last experiment, there were
20 normal nodes and 20 malicious nodes. In this experiment
the transmission rate of malicious nodes was increased to 35
Kbps. In each experiment the target of the flows was gateway
node.
Fig.7. Throughput of UDP Flows

Packets
Normal Flows
LDFA Flows
HDFA Flows
Time (Sec)
Fig.8. Throughput of TCP Flows
Furthermore, it is evident from Fig. (9, 10) that with

increasing the flows transmission rates in each experiment,
packet dropping rate is also increased.
Packets
HDFA- Pkt.Drop.Rate
LDFA-Pkt.Drop.Rate
Normal-Pkt.Drop.Rate
Time (Sec)
Fig.9. Packet Drops of UDP flows
124
Proceedings
Engineering & Technology (IJARCET), vol. 2, no. 5, pp. 17661770, 2013.
Packets
HDFA- Pkt.Drop.Rate
[8] D. J. Jingle and E. B. Rajsingh, "Defending IP spoofing attack

and TCP SYN flooding attack in next generation multi-hop
wireless networks," International Journal of Information &
Network Security (IJINS), vol. 2, no. 2, pp. 160-166, 2013.
LDFA-Pkt.Drop.Rate
Normal-Pkt.Drop.Rate
[9] Thwe Thwe Oo and Thandar Phyu, "DDoS detection system

based on a combined data mining approach," 4th International
Conference on Science and Engineering, pp. 315-319, 2013.
Time (Sec)
[10] G. Wang, J. Hao, J. Ma, and L. Huang, "A new approach to

intrusion detection using artificial neural," Expert Systems with
Applications, vol. 37, no. 9, pp. 6225-6232, 2010.
Fig.10. Packet Drops of UDP flows
IV. CONCLUSION AND FUTURE WORK
[11] S. Khan and K. K Loo, "Real-time cross-layer design for largescale flood detection and attack traceback," Network Security,
vol. 23, no. 5, pp. 9-16, 2009.
In WMNs, mesh gateways are mostly susceptible to

DDOS attacks. Numerous techniques are proposed to detect
distributed flooding attack in WMNs, but still there is a need
of an intelligent security mechanism to avoid the influence of
intruders on WMNs. An intelligent mechanism based on
ANN, is being proposed in this paper to classify incoming
flows at gateway into three different categories. Simulations
results verified that DFAD is efficient and precise in
classifying the incoming flows. In future, DFAD will
differentiate flash crowd form attack flows. Moreover, we will
also identify and trace back source of the ongoing distributed
flooding attack.
[12] D. Novikov, R. V. Yampolskiy, and L. Reznik, "Artificial

intelligence approaches for intrusion," in IEEE Long Island
Systems, Applications and Technology Conference (LISAT
2006), Long Island, NY, 2006.
[13] E. Gelenbe, "Learning in the recurrent random neural network,"
Neural Computation, vol. 5, no. 1, pp. 154164, 1993.
[14] S. M. A. Burney, M. S. A. Khan, and Dr. T. A. Jilani, "Feature
deduction and ensemble design of parallel neural networks for
intrusion detection system," International Journal of Computer
Science and Network Security (IJCSNS), vol. 10, no. 10, pp.
259-267, 2010.
[15] G. OKE and G. Loukas, "A denial of service detector based on
maximum likelihood detection and the random neural network,"
The Computer Journal, vol. 50, no. 6, pp. 717-727, 2007.
ACKNOWLEDGEMENT
The authors extend their gratitude to the Research Centre,
Institute of IT, Kohat University of Science & Technology and
the Higher Education Commission of Pakistan, for funding
this research work.
[16] B. B. Gupta and R. C. Misra,M. Joshi, "ANN based scheme to

predict number of zombies in a DDOS attack," International
Journal of Network Security, vol. 14, no. 2, pp. 61-70, 2012.
[17] The network simulator - ns-2 web page,"http://nsnam.isi.edu/
nsnam/index.php/Main_Page".
REFERENCES
[1] S. Khan, K. K. Loo, and Z. Din, "Framework for intrusion
detection in IEEE 802.11 wireless mesh," International Arab
Journal of Information Technology, vol. 7, no. 4, pp. 435-439,
2010.
[2] D. Benyamina, A. Hafid, and M Gendreau, "Wireless mesh
networks design a survey," IEEE Communications Surveys &
Tutorials, vol. 14, no. 2, pp. 299-310, 2011.
[3] F. Xing and W. Wang, "Understanding dynamic denial of
service attack in mobile ad hoc networks," in IEEE Military
communication conference (MILCOM), 2006.
[4] H. Beitollahi and G. Deconinck, "Analyzing well-known
countermeasures against distributed denial of service attacks,"
Computer Communications, vol. 35, no. 11, pp. 13121332,
2012.
[5] B B Gupta, Joshi, and Manoj , "Distributed denial of service
prevention echniques," International Journal of Computer and
Electrical Engineering, vol. 2, no. 2, pp. 268-276, 2010.
[6] K Reyhaneh and F Ahmad, "An anomaly-based method for ddos
attacks detection using rbf neural networks," International
Conference on Network and Electronics Engineering IPCST,
vol. 11, pp. 44-48, 2011.
[7] Thwe Thwe Oo and Thandar Phyu, "A statistical approach to
classify and identify DDoS Attacks using UCLA dataset,"
International Journal of Advanced Research in Computer
125
Proceedings
Touch Panel Based Modern Restaurants Automation

using Zigbee Technology
1,*
Aamir Nawaz,2Faiz Jalil
Institute of Engineering & Technology

Gomal University
DIKhan, Pakistan
1
engr.aamir09@gu.edu.pk
2
fj_kulachvi@yahoo.com
Abstract
In present scenario, automation is in high demand.

Management of present restaurants and hotels are much
dependent on manpower. This project deals with automation
of restaurants with wireless touch panel based menu systems.
This project reduces manpower required for taking order from
customers. Full menu of all restaurants eatable items are
displayed on touch panel for selection. Customers have to put
their orders through touch panels which will be received in
kitchen without any interaction of waiters. Zigbee has been
used for wireless link between touch panels of kitchen and
restaurant tables. PIC microcontroller has been used for
coding of menu display on touch panel and transmitting and
receiving on zigbee modules. Project has been implemented
and tested in software such as Proteus 8. Furthermore,
hardware implementation has been done on PCB layout. By
using embedded system with wireless Technology whole
procedure will built. The order from selected table number go
to server then serve items to selected table customer .We can
speed up the serving and attract customer with non volatile
memory capable of retaining data for over ten years .By
dumping the code in microcontroller ISP is used
Index TermsZigbee, Restaurant automation, Micrcontroller
I. INTRODUCTION
Restaurant business is one of the most profitable
businesses. Therefore, the importance of food serving is of
great significance. Over the years, food and the relative job of
serving has grown so much that need for facilitation and
automation has been increased. industries. Nowadays, when
profit is the prime concern and its measurement has increased
from million to billions, every bit is done to increase profit.
We have made an effort to bring technology into the dining
menu of customers. This research aims to not only improve the
business of restaurants but also to incorporate the essence of
science in dining menu.
Arun in [1], implemented a train anti-collision and level
crossing system based on zigbee technology. He has created
four modules which are train system module, control center
system module, signaling post system module and level

crossing at gate module. He has simulated proposed system in
Proteus electronic simulation package.
Ramasamy [2] has proposed a fire avoiding system on
running train. In his proposed system, train driver will know in
case of fire in any section of train and stop it instantly to take
necessary action.
In [3], Blaho proposed a zigbee wireless network to control
speed and other parameters of train model. He has created mfiles and Simulink environment which can be used for teaching
purposes.
In [4], Vinodhini proposed a restaurant system for two way
ordering system. He has provided table of customer with touch
screen and kitchen with liquid color display (LCD). He has
used ARM cortex M3 for simulation.
In [5], Wang has introduced zigbee technology for
establishing wireless sensor networks (WSN) and its
applications in industry and home appliances.
Bhargave in [6], has introduced a system for restaurant
using android technology. He proposed integration of Order
System Kitchen Order Ticket (KOT), Billing System and
Customer Relationship Management System (CRM).
Implementation of such integration has increased quality and
speed of restaurant service.
Wahab in [7], has presented a new development in
restaurant ordering system. He built up network system
integration for connecting customer table with management
panel. He used microcontroller with VB and RS 232
communication port. Proposed system performed well in
network based system.
The proposed system will have a complete Touch Panel
device in restaurants and hotels that will help to reduce the
human interaction and save time. This project implements a
system that is on customer table by having touch panel menu
ordering system. We have one touch technology by selecting
the items from menu list display in LCD on customer table
The basic concept of the project is to reduce human effort
for the customers and also to deliver them with best facilities.
Wireless communication link between customers table and
kitchen has been established for placing order on customers
side and verifying it on kitchen side. Furthermore, bill details
126
are also displayed on customers screen. Management can

change menu by editing it and adding more items or changing
cost of each item as per market rates.
This paper is organized as explained further. Section 2
contains problem statement. Section 3 contains working of
Zigbee technology. Section 4 contains overview of proposed
system for restaurants. Section 5 contains comparison of
proposed system with previous systems. Section 6 explains
simulation and results. Section 7 contains conclusion of this
paper.
II. PROBLEM STATEMENT
The operational system is very tedious .Manual order
placements with the help of waiters have been carried out in
restaurants. Menu card are normally presented on to the
customer out of whom he can select available eatable items.
All this process is cumbersome for customers from order
placement till bill payment.
Moreover, changing of menu cards or adding new items to
menu, causes problems for responsible persons to update menu
cards frequently. Therefore, existing system is time taking and
requires a lot of human effort. The existing system is noncomputerized. Such manual systems lead to problems because
the waiter might not understand what the customer had ordered
therefore serving him/her a different menu. This could be quite
embarrassing as it will lead to customer dissatisfaction and ruin
of restaurant reputation.
III. ZIGBEE TECHNOLOGY
ZigBee Module is a low-cost, low-power, wireless mesh
networking standard. With minimum amount the technology to
be widely deployed in wireless control and monitoring
applications, the low power-usage allows longer life with
smaller batteries, and the mesh networking provides high
reliability and larger range
This network device having qualities of electric powersaving, reliability, low cost,large capacity and security, and it
has very wide application in various fields of automatic
control. The target application domains are aimed at industry,
home automation, telemetry and remote control
Technical Data
o Use Industry Standard for environment
o Maximum Transmission Distance:
o Internodes Barrier Free: 200 meters( 658feet)
Wireless Frequency:
2.4G ISM License-free Frequency Band
Channel Mode:
16 Channels Can Be Specified or the Best Channel Can Be
Automatically Selected
Antenna Configuration:
Built-in 2.4G Ceramics Antenna
Node Type:
Center Node, Routing Node, Terminal Node, or Software
Set
Serial Rate:
1200-115200
Proceedings
Send Mode:
Broadcast Send or Destination Address Send
Working Voltage
DC-3.3V
Peak Current
40mA
Best features is Low power consumption. At the low power
consumption standby mode, two No.5 dry-charged batteries
can support one node to work 6 ~ 24 months, or even longer.
Low cost. Because of dramatically simplifying the protocol. It
is protocol patent fee, free and low rate. Zigbee can work at the
low rate of 20 ~ 250 kbps. Short range transmission range is
generally between 10 ~ 200 m ( 32.8feet~ 656feet) .Short time
delay. The response speed of Zigbee is very fast, in general, it
merely need 10ms from the into work state and it merely need
20 ms from nodes connect into the network state.
For high capacity Zigbee can adopt star topology, tree
topology and mesh network structure, composing of up to 65,
000 network node. For high security. Zigbee provides a threetier security model, including the use of Worry-Free Security
settings, the access control list ( ACL) to prevent illegally
accessing the data and Advanced Encryption Standard ( AES
128) symmetry password. For License-free frequency band.
Zigbee adopts direct sequence spread spectrum of the industrial
scientific medical (ISM) band. 2. 4 GHz (global).
Sensor Network (WSN) Automatic meter reading system of
water, gas, heat, electricity meters Intellectual traffic control,
signal lights control and street lights control Fire safety alarm,
building monitoring Catering order, canteen s sale of food
system Access control, time and attendance system PTZ
monitoring, engine room equipment monitoring Store &
Logistics, laser guns, bar code reader RFID remote data
transmission Meeting ballot system
IV. OVERVIEW OF PROPOSED SYSTEM
The proposed system has been developed for making
restaurant ordering system easy. This system is also capable of
keeping record of customers orders.
The proposed systems support restaurant operation by
performing following tasks:
To put, view or change submitted order on the customers
table module.
To display menu on customers module interface.
To allow kitchens module to give feedback about placed
order of customer (whether ordered items are available).
The proposed system will handle efficiently all problems
faced by previous manual systems. Furthermore, system will
store and analyze information, which will be provided by
customer and kitchen staff and perform task automatically.
The proposed system has also following added features that
can increase productivity of the system:
Data handling with ease and accuracy
Reduction in paper work done by restaurant waiters or
managers.
127
A. Zigbee transceiver at customer end (on table)

. The customer module is play key role our project since the
customers are the vital aspect for any business firm. The
customer module consists of sensitive touch hardware. Our
project was implemented using FRIENDLYARM. The best
features of the hardware includes impressive menu, comprising
of self-generated item number, item name, item description, an
interesting feature is the item picture, tax, price and quantity
selection checkbox allowing a maximum level of 10. After
selecting the item the customer selects the add order button
upon which the page refreshes and when he clicks the view
order button he can see the item he has ordered along with the
cost and if the customer wishes he can edit the order and start
ordering afresh otherwise he can confirm the order by pressing
the place order button, and a message will come and after
taking food the customer can also send his valuable.
Figure 1: Block diagram of customer end module

1) Implementation
Microcontroller pic 16F877A is place in middle and all
components are connected with this microcontroller firstly
connected with DC supply interfacing via 10k resister five pins
are connected with lcd and eight pins are interfacing with touch
pad and three pins are connected with zigbee module and eight
pins are first connected with 10 k resisters then coneected with
led and these interfacing are with red lines to short these
components
Proceedings
Figure 2: Customer end module operating

B. Zigbee transceiver at restaurant management end (in
kitchen)
As soon as the customer confirms his order, the order is
displayed in the Central Kitchen . The order comprises of the
table number, order number and dish name. The status of
incoming order is made pending and as soon as the chef
prepares the food he can set the status as finished. The
importance of this module is that it plays a vital part in time
reducibility.
Figure 3: Block diagram of kitchen end module

1) Implementation
After deep studying and analysis of these components
interfacing one by one to gain the desire results first available
all components and microcontroller is the main in which all
other components interfaced
As microcontroller is placed and given the power supply
with two pins and zigbee module is connected with three pins
and eight pins are connected with LCD and some pins are
connected with resistors of 10 k and then interfacing with led
different is interfacing three relays on this side with resistors
that if some other components are placed with this u can easily
attach like ATM dipper etc.
128
Proceedings
V. COMPARISON OF PROPOSED SYSTEM WITH PREVIOUS

SYSTEMS
Table 1 shows comparison results on the basis of different

features considered for comparison. For comparison, Pixel
point [9], LRS[10], Eric software[8] and Smart order system[7]
are considered. As it can be seen that proposed system
outperforms all other systems in comparison.
Figure 4: Kitchen end module operational

TABLE 1. COMPARISON OF PROPOSED SYSTEM WITH PREVIOUS SYSTEMS
Features
Graphics
ATM Sweeper
Prices
Group Orders
Status of Ordering
Wireless Network
Touch Screen
Pixel point
NO
NO
NO
NO
NO
YES
YES
LRS
NO
NO
NO
NO
NO
YES
NO
Eric soft
NO
NO
NO
NO
NO
YES
YES
VI. SIMULATION AND RESULT EXPLORATION

In restaurant, customers table module has a touch screen
display in which menu is shown. When this menu is pressed,
full menu of restaurant eatables are shown on screen to be
selected. Once order is selected and forwarded to kitchen, it
uses zigbee transmitter and receiver to deliver that information
to kitchen. Kitchen module will receive order which will be
displayed on LCD. Kitchen person will confirm order by
pressing confirm button or reject any unavailable eatable items
from menu. Then, customer will get information regarding
unavailability of any items and can re-order. Furthermore,
waiter can be called by pressing Call waiter button.
The proposed system can be used with line following
robots or conveyer belts to deliver order till customers. This
will reduce more human effort as there will be no waiter
required to deliver order to tables. Other benefits are:
Provision of facility for handling text electronically using
powerful processors to produce elegant and error free
documents.
In addition to storing the organizations operational data
on disk drive for back up.
With the installed software, product ordering and delivery
was made easier.
Smart order system

NO
NO
NO
YES
NO
NO
NO
Proposed System
YES
YES
YES
YES
YES
YES
YES
The systematic approaches used during each phase of the

software development provides a clear road map that
would be of immense help to anyone carrying out research
work in this area.
VII. CONCLUSION
This research contributes to recent working environment of

restaurants and can change it from previous manual system to
an automated restaurant with a lot of facilities to customers as
well as restaurant management. Furthermore, it will be helpful
in attracting customers to see and avail such facilities. Zigbee ,
as a wireless link between customers table and kitchen can be
replaced by any other wireless technologies for larger
restaurants which supersedes zigbees range of communication.
Line following robots or conveyer belts can be used to deliver
eatable items to customers instead of waiters.
REFERENCES
[1] Arun P., Sabarinath G., Implementation of ZigBee based Train
Anti -Collision And Level Crossing Protection System for
Indian Railways." Int. J. of Lat. Trends in Eng. and Tech., Vol.2,
pp: 12-18, January, 2013.
129
[2] R. P. Ramasamay, M. P. Kumar, S. S. Kumar, R. R. Raman,

Avoidance of fire accident on running Train using ZigBee
Wireless Sensor Network, Int. J. of Inform. and Comp. Tech.,
Vol.3, pp: 583-592, 2013.
[3] M. Blaho, J. Ozimy, M. Ernek, M. Foltin, Zigbee
implementation in Railroad model.
[4] B. Vinodhini, K. Abinaya, R. Roja, M. Rajeshwari, Wireless
Two-way Restaurant Ordering System via Touch Screen,
Vol.3, pp: 01-05, 2014.
[5] W. Wang, G. He, J. Wan, Research on Zigbee wireless
communication technology, Int. Conf. on Elect. and Cont. Eng.
(ICECE), 2011.
[6] Bhargave, A., Jadhav, N., Joshi, A., Oke, P., & Lahane, M. S.
(2013). Digital Ordering System for Restaurant Using
Android, International Journal of Scientific and Research
Publication, Vol.3, 2013.
Proceedings
[7] Wahab, M. H. A., Kadir, H. A., Ahmad, N., Mutalib, A. A., &
Mohsin, M. F. M. (2008, August). Implementation of networkbased smart order system, International Symposium on
Information Technology, 2008. ITSim 2008. (Vol. 1, pp. 1-7).
IEEE.
[8] .[8] M.H.A. Wahab, H.A. Kadir, N. Ahmad, A.A. Mutalib and
M.F.M. Mohsin, Implementation of network-based smart order
system, International Symposium on Information Technology
2008
[9] PAR PixelPoint PixelPoint POS Brochure [Accessed: 11 May
2011]
[10] Advanced Analytical, Inc (October 2004) LRS Restaurant
Server Pager, [Accessed: 11 May 2011]
130
Proceedings
Introducing space plant biology to students through

hands-on activities using clinostat
Ferheen Ayaz
Qurat-ul-Ain

Karachi, Pakistan
ferheen@msn.com

Karachi, Pakistan
ain_haider@hotmail.com
AbstractSpace biology is an important topic for research

and education in order to explore future life in space. Plant
growth and development largely depends upon gravitational
acceleration on earth. The growth process of a plant in space is
significantly different from that on earth due to the absence of
gravity, which is referred as zero gravity or microgravity. Oneaxis clinostat simulates microgravity environment on earth. On a
horizontal clinostat, an object is rotated perpendicular to the
force of gravity. Although the gravity is present on an object
rotated by clinostat, its continuous reorientation nullifies the net
effect of gravitational acceleration. One-axis/horizontal clinostat
is a useful tool for space education. The simulated microgravity
created by the clinostat, despite its limitations, causes significant
changes in plant growth which may lead to introduce school and
college level students about plant life in space and creating their
interest in space biology. This paper describes some of the
experiments conducted on one axis clinostat, by a group of
teachers and students of SUPARCO Institute of Technical
Training, which proved to be a fruitful activity resulting in
learning of gravitropism basics of plants and seeds. The article is
aimed towards helping the educators in creating hands-on
learning activities for students to have an idea about life in space
by making use of facilities available on earth. The orientation of
roots grown from radish seeds (Raphanus sativus), area and
length of roots and shoots grown from pea seedlings (Pisum
sativum), length and bending angle of shoots of Marigold plant
(Tagetes patula) and growth rate and direction of wheatgrass
(Thinopyrum intermedium) were observed on the clinostat. The
roots of radish seeds did not grow downwards in the direction of
gravity but oriented themselves at random angles on the
clinostat. Pea seedlings also showed different growth rates on
clinostat as compared to their growth on earth. The experiment
on Marigold plant resulted in faster growth in length and larger
bending angle of clinorotated plant shoots as compared with
stationary plant shoots. The growth rate of wheatgrass on earth
and on clinostat was nearly same but their bending angles were
different in both conditions. All the experiments provoked the
minds of students to make inferences on seeds and plants growth
in space. The experiments described in the paper create
inquisitiveness among teachers and students to observe many
other parameters in plant growth under microgravity.
Keywordsclinostat; gravitropism; microgravity
important concerns is to analyze the development of plants in

space. The reason behind this concern is the dependence of
plants and animals upon each other for their survival and
flourishing of ecosystem [1]. Study of plant growth in space is
different from conventional plant growth on earth mainly
because of the fact that it is greatly influenced by gravitational
acceleration. Recent experiments in space have resulted in
modified growth rates of roots and shoots of different plants
and the major reason of this modification is the absence of
gravity in space [2].
In order to promote space exploration related activities and
establish plant life in space, one of the key initiatives is to
promote space education among young students. Creating
interest of students in space sciences may lead to increase their
curiosity towards exploring life support systems in space and
planetary surfaces. Due to limited facilities of space research
in underdeveloped countries, students are least interested in
space biology. The students are less inclined to study space
sciences because of unavailability of resources and
opportunities to do practical work in this field. The practical
work for science students plays a very important role in their
learning [3]. As it is difficult to create practical setups of
complete space environment on earth, establishing a simulated
microgravity environment can provide excellent practical
opportunity for researchers and students to study life sciences
and plant and animal life in space [4].
Fig. 1. One-axis clinostat to simulate microgravity on earth.
Microgravity for a very short period of time can be created

on earth by elevator, drop tower, rocket and aircraft [5] but
changes in plant growth cannot be observed by these methods
as their growth process is comparatively slower. The efficient
instruments for conducting biology experiments in simulated
I. INTRODUCTION
Considering the long term goals of space agencies and
seeking habitation of extraterrestrial surfaces; one of the
131
microgravity environment are 2D and 3D clinostats [6].

Keeping in view the practical needs of space education,
SUPARCO (Space and Upper Atmosphere Research
Commission) Institute of Technical Training acquired 2D
clinostat, as shown in Fig. 1, under ZGIP (Zero Gravity
Instrument Project) initiated by United Nations Human Space
Technology Initiative (HSTI) in 2013. A 2D or one axis
clinostat has one rotational axis which rotates perpendicular to
the direction of gravity vector in order to create the effect of
microgravity and is suitable for studying the behavior of
germinating seeds and analyzing plant growth [7]. Although
horizontal clinostat does not completely create zero gravity
environment on the object under study [8], it can still be used
for educational purposes to demonstrate changes in plant
growth by alteration of gravity.
Proceedings
Fig. 2 (a). First Petri dish with roots parallel to gravity vector.
This article describes some experiments conducted on

seeds and plants using clinostat at SUPARCO Institute of
Technical Training by a group of 2 instructors and 10 students
aged between 17 to 19 years. The objective of describing the
experiments is to promote science teachers and students to
study space biology practically and to provoke the minds of
interested learners to bring up similar but new ideas to
experiment and discover effects of gravity on plants.
The organization of the rest of the paper is as follows:
Section II describes the experiments performed using
clinostat, Section III discusses results of the experiments and
conclusion is mentioned in Section IV.
II.
EXPERIMENTS PERFORMED USING CLINOSTAT
A. The radish seeds

The first experiment to study gravitropism on the roots of
radish seeds (Raphanus sativus) was conducted as described in
[7]. Nine samples of radish seeds were placed on each of the
three Petri dishes. The Petri dishes were half filled with agaragar solution to aggravate the process of germination.
Humidity, temperature and light conditions were kept same for
each Petri dish. After 48 hours, when the roots started to grow
downward, first Petri dish was placed vertically with roots
along the direction of gravity, second was turned 90o so that
the roots were held perpendicular to the gravity vector and the
third was attached to clinostat using double sided tape. The
rotation disc of clinostat was held horizontally. The placement
of Perti dishes is illustrated in Fig. 2. The clinostat was set to
rotate with a speed of 10 r.p.m. for 4 hours and photographs of
each Petri dish were captured using a digital camera. After the
experiment, the photographs were observed using Image J
software. The orientation of roots was analyzed in terms of
angle created as the roots bent during the experiment.
Fig. 2 (b). Second Petri dish with roots perpendicular to gravity vector.
Fig. 2 (c). Third Petri dish attached to clinostat.
B. The pea seedlings

This experiment was designed by the working group to
observe not only the roots, but also the shoots grown from pea
seeds (Pisum sativum). Four test tubes were half filled with the
agar-agar solution made by the procedure defined in [7]. A pea
seed was placed inside each test tube such that it lies exactly
in the middle of the tube as shown in Fig. 3. After 7 days, the
shoots and roots of the pea seeds were grown. Two test tubes
were placed horizontally perpendicular to the gravity vector
and the other two test tubes were attached to clinostat. The
132
clinostat was set to rotate on horizontal axis at 15 r.p.m. for 72

hours in a dark room. The area and length of both root and
shoot of each pea seedling was observed at the end of
experiment.
Proceedings
were soaked in water overnight and then grown in two Petri

dishes filled with soil for 10 days under direct sunlight and
normal gravity conditions of earth. When the grass grew out of
the Petri dishes, one of the dishes was placed on earth with
grass growing in the direction opposite to the gravity and the
other was attached with the clinostat, as shown in Fig. 5. The
clinostat was set to rotate on horizontal axis at 10 r.p.m. for 4
days in a room with a 100 Watt yellow light bulb turned on
continuously. The photographs of both Petri dishes were
captured each day. The soil in Petri dishes was watered each
day using a spray bottle. At the end of experiment, the length
and bending angle of some samples of wheatgrass were
observed in the photographs using Image J software.
Fig. 5. Experimental setup for observing growth of wheatgrass under normal

gravity on earth and simulated microgravity produced by clinostat.
Fig. 3. Placement of pea seed in test tube.
C. The Marigold plant

The growth of shoots of Marigold plant (Tagetes patula) in
simulated microgravity produced by clinostat was compared
with the growth in normal gravity conditions. In this
experiment, six small Marigold plants were observed. Initially,
the length of the shoot of each plant was measured. The stripes
with the interval of 5mm were made on each complete shoot
using permanent marker. Six test tubes were filled with water
and a Marigold plant was inserted in each test tube such that
roots were dipped inside water and shoots appeared from just
outside the tube as shown in Fig. 4. Three test tubes were
placed horizontally perpendicular to the gravity vector and
other three test tubes were attached to clinostat. The clinostat
was set to rotate on horizontal axis at 5 r.p.m. for 4 hours in a
dark room. The length and bending of each shoot was
observed at the end of experiment.
III. RESULTS
A. The radish seeds
The results of the experiment were as follows:
The roots in first Petri dish grew in the direction of
gravity and the effect is called positive gravitropism,
therefore their bending angle was approximately zero
degree.
Fig. 6 shows that the bending angle of second Petri
dish (90-degree turned) was gradually decreasing
which shows their growth towards the direction of
gravity with the passage of time. However, the decline
in the angle was not at a constant rate.
Fig. 6 does not show a regular pattern in bending angle

of third Petri dish (clinorotated) and hence we can
conclude that in microgravity, the roots may grow in
any random direction and not necessarily downwards.
Summary of the result is presented in Table I.
Fig. 4. Placement of Marigold plant in test tube.
D. The wheatgrass
The fact that wheat is one of the major food crops of
Pakistan motivated the working group to observe its growth
behavior on clinostat. Wheat seeds (Thinopyrum intermedium)
133
Fig. 6. Bending angle of roots of radish seeds on 90 o turned and clinorotated

Petri dish with respect to time.
TABLE I.
BENDING ANGLE OF ROOTS OF RADISH SEEDS

Time
(minutes)
Proceedings
C. The Marigold plant

On an average, the shoots of Marigold plant grew more in
4 hours on clinostat as compared to earth. Table III and the
graph in Fig. 7 summarize the result. The following important
points were resulted:
The average growth of shoots on earth was 0.034 cm
and on clinostat was 0.305 cm.
Average angle of nine

seeds (degrees)
90o turned
Clinorotated
35.3897
9.241
30
28.2403
14.584
60
22.9933
15.775
90
26.0908
15.798
120
24.8322
14.128
150
25.962
6.935
180
24.0473
12.757
210
21.8095
15.607
Bending of shoots was observed both on clinostat and

on earth, but the shoots on clinostat bent at larger
angles than the shoots which were placed horizontally
on earth.
TABLE III.
Gravity
condition
1-g
B. The pea seedlings

The results of the pea seedlings grown for 3 days on
horizontal clinostat at 15 rpm compared with stationary
seedlings were as follows:
Roots from clinorotated plants had a smaller cross
sectional area compared to the stationary control with a
difference of -3% whereas shoots from clinorotated
plants had a larger diameter relative to the stationary
control with a difference of 23%.
Simulated
microgravity by
clinostat
GROWTH OF SHOOTS OF MARIGOLD PLANT
Test
tube
Initial
length
of
shoot
(cm)
L1
5.3
Final
length
of
shoot
(after 4
hours)
(cm)
L2
5.454
5.6
6.3
Difference
(cm)
L2-L1
Angle of
bending
after four
hours
(degrees)
0.154
-0.256
5.602
0.002
11.911
6.385
0.085
1.193
5.4
5.655
0.255
28.944
4.9
5.428
0.528
30.42
5.6
5.733
0.133
30.379
Clinorotated pea seedlings showed an increase in root

length by 3% and increase in shoot length by 27% as
compared to stationary control (1-g on earth).
Summary of the result is presented in Table II.
TABLE II.
Fig. 7(a). Growth in length of shoots of Marigold plant.
RESULT OF EXPERIMENT ON PEA SEEDLINGS

Stationary
Individual
test tubes
Clinorotated
Average
Individual
test tubes
Average
Roots
Area
(mm
square)
Length
(mm)
0.046
0.041
7.56
6.791
0.048
0.0435
7.1755
0.036
7.91
6.813
0.042
7.3615
Shoots
Area
(mm
square)
Length
(mm)
0.046
0.023
7.507
3.793
0.0345
5.65
0.061
0.024
9.919
4.48
Fig. 7(b). Bending angles of shoots of Marigold plant.

0.0425
D. The wheatgrass
The results of the experiment are as follows:
7.1995
As shown in Fig. 8(a), the length of wheatgrass showed

similar growth rate on earth and on clinostat and it was
decreasing per day. Calculated growth rate on earth
134
was 0.118 cm/hour from Day1 to Day2, 0.026 cm/hour

from Day2 to Day3 and 0.024 cm/hour from Day 3 to
Day4. It was almost the same on clinostat. The
calculations came out to be 0.112 cm/hour from Day1
to Day2, 0.027 cm/hour from Day2 to Day3 and 0.026
cm/hour from Day 3 to Day4.
As shown in Fig. 8(b); on earth, the wheatgrass grew
almost straight and angles with respect to the direction
of gravity were in the range of 0 to 3o, showing
almost no bending. On clinostat, the angles were
varying in the range to 10o to 14o which is the evident
proof of bending.
IV. CONCLUSION
Four experiments on different seeds and plants were
conducted on clinostat. The results show that all plants do not
exhibit similar behavior in microgravity. Absence of gravity
speeded up the growth rate in pea seedlings and Marigold
plant but did not influence wheatgrass. The diversity of the
experiments described in this article encourages the students
to try out experiments on other plants and seeds available to
them. Various other parameters like temperature, humidity,
type of soil, amount of water and nutrients provided can also
be modified to discover new effects of gravity and
microgravity on plants.
Summary of the result is presented in Table IV.

TABLE IV.
Acknowledgment
The study is carried out as a part of Zero Gravity
Instrument Project initiated by United Nations Office for
Outer Space Affairs (UNOOSA). We gratefully acknowledge
UNOOSA for donating clinostat, Training Division, HR Dte
Gen, Pakistan Space and Upper Atmosphere Research
Commission for their coordination in publishing the article
and Horticulture Cell, Pakistan Space and Upper Atmosphere
Research Commission for assisting in plant growth.
LENGTH AND ANGLE OF WHEATGRASS DRUING 5 DAYS OF

EXPERIMENT
Measurement
of:
Average Length (cm)
Average Angle
(degrees)
On
On earth
clinostat
Day
On earth
On
clinostat
3.69
2.30
0.931
13.73
6.54
5.00
-0.69
8.77
7.19
5.67
2.81
10.80
7.77
6.31
-2.44
8.39
Proceedings
References
[1]
[2]
[3]
[4]
[5]
Fig. 8(a). Length of wheatgrass per day on earth and on clinostat.
[6]
[7]
[8]
Fig. 8(b). Angle of wheatgrass per day on earth and on clinostat.
Sponsored by Pakistan Space and Upper Atmosphere Research

Commission.
135
R. Ferl, R. Wheeler, H. G. Levine, and A. L. Paul, Plants in space.

Current opinion in plant biology, vol. 5, no. 3, 2002, pp. 258-263.
T. Hoson, Plant growth and morphogenesis under different gravity
conditions: Relevance to plant life in space, Life, vol. 4, no. 2, 2014,
pp. 205-216.
V. N. Lunetta, A. Hofstein, and, M. P. Clough, Learning and teaching
in the school science laboratory: An analysis of research, theory, and
practice, Handbook of research on science education, 2007, pp. 393
441.
A. Niu, M. Ochiai, and H. J. Haubold, The United Nations Human
Space Technology Initiative (HSTI): Science Activities, IAC-12A2.5.11, 2012.
M. J. Rogers, G. L. Vogt, and M. J. Wargo, Microgravity: A Teacher's
Guide with Activities in Science, Mathematics, and Technology,
NASA, EG-1997-08-110-HQ, 1997.
M. Cogli, The fast rotating clinostat: a history of its use in gravitational
biology and a comparison of ground-based and flight experiment
results, Gravitational and Space Research, vol. 5, no. 2, 2007.
Programme on Space Applications: Teachers Guide to Plant
Experiments in Microgravity, United Nations Office for Outer Space
Affairs, 2013.
A.H. Brown, O.A. Dahl, and D.K. Chapman, "Limitation on the use of
the horizontal clinostat as a gravity compensator," Plant Physiology, vol
58, no. 2, pp. 127-130, 1976.
Proceedings
The Political and Economic Feasibility of Current

Space Resource Management Policies
Faizan Muhammad
Muhammad bin Munim
Lahore, Pakistan
faizan.muhammad7@outlook.com
Lahore,Pakistan
muhammadbinmunim@gmail.com
AbstractAt the rapid pace with which the interest in

utilization of extra-terrestrial resources is developing, it would be
no surprise if such projects are practically begun by the next
decade. This, however, poses a unique dilemma to the scientists
and politicians alike. There is no global consensus as yet
regarding the specific rules and regulations that would come into
play, nor is there any international framework to establish a legal
foothold in the matter.
To further complicate matters, not only are national
governmental space agencies interested in such projects, but
private companies too wish to take part in the process. Countries,
such as the USA for example, have a relatively elaborate
procedure for private companies but many others do not.
The 1967 Treaty on Principles Governing the Activities of
States in the Exploration and Use of Outer Space, including the
Moon and Other Celestial Bodies by the United Nations was
successful but the 1979 Agreement Governing the Activities of
States on the Moon and Other Celestial Bodies turned out to be
a disappointment. The flaws that caused it to be treated that way
were identified and addressed.
The feasibility and possible consequences of the
aforementioned treaties and legislature were analyzed with
respect to future plans of government and private agencies as
well as relevant historic parallels. Certain limitations were
exposed in the current system which may lead to escalated
controversies in the future. Suggestions have also been made to
improve or redesign certain aspects of existing systems to work
together efficiently and smoothly all over the globe.
Suleman Saleem
Lahore, Pakistan
msulemansaleem727@gmail.com
Moon Treaty calls for solidification the management of Space

under an international regime. It provides equal rights to all the
states in terms of conducting research in space. Furthermore it
bans commercial activity in space, giving the power of
resource extraction and allocation to international regime
approved parties only. Another distinction of the Moon Treaty
is that unlike the Outer Space treaty, it also prohibits ownership
of any part of the celestial bodies by private parties not
necessarily tied to any nation.
In the following sections, the incentives and proposed plans
of various countries and organizations regarding lunar and
asteroid mining will be discussed as case studies in order to
put emphasis on how current treaties governing space
exploration and resource management are not feasible.
To impress upon the urgency of the matter, the reasons for
controversies will be discussed, leading to the importance of
treaties in the international landscape and through the analysis
of treaties and treaty organizations a parallel will be drawn
between the historic treaties and conflict and the current space
situation to show that the current treaties governing space
exploration and resources management solutions are not
viable.
Finally some suggestions will be made to improve the state
of current space resource management policies to help avoid
any potential future conflicts.
LUNAR RESOURCES
A. Helium-3 Extraction
Index TermsSpace Econopolitics, Space Ownership, Space

Law
TABLE 1.PROPERTIES AND MAJOR APPLICATIONS OF HELIUM-3
INTRODUCTION
Due to the fact that the Moon Treaty has not been ratified
by any nation that has achieved spaceflight (and will thus play
an active role in space expeditions in the near future), the Outer
Space Treaty of 1967 holds greater importance than the Moon
Treaty and hence currently forms the legal framework of
international space law . The treaty prohibits any military
activity in space, confining the use of the moon, its orbits, and
other celestial bodies to peaceful purposes only. Furthermore,
the treaty explicitly forbids any claim of sovereignty by any
nation by any means whatsoever.
The clauses of the Moon Treaty are very similar to those of
the Outer Space Treaty, the major difference being that the
136
Property
Application
Does not solidify, even

at 0K, at atmospheric
pressure
Used for Liquid cooling

of Low Temperature
Superconductors
Second lightest element,

after Hydrogen
Used as a substitute for

Hydrogen
in filling
balloons
Helium is the element

with
the
smallest
molecular size
Used for Leak detection
Has
extremely
solubility in water
low
Used as a diving gas
Helium-3
radioactive
not
Used as a heat transfer

medium in fusion reactors
is
The importance of the noble gas being realized, it is

important to mention that there is a shortage of Helium on
Earth. Currently, Helium supplies are obtained from the
extraction of natural gas. The United States produces about
75% of the worlds Helium and roughly 30% of the worlds
Helium supply comes from the US federal Helium reserve. The
problem is that Helium reserves are being depleted at a rapid
rate because of wide use of the gas. Walter Nelson, director of
helium sourcing for Air Products and Chemicals Inc. predicts
the usable life of the US reserves to last till 2018-2020. This
will have a global impact, especially on healthcare and smallscale scientific research. MRI machines need Helium to
operate and they will become difficult to maintain with a
shortage of Helium, ultimately making it very difficult to
obtain MRIs.
Furthermore, Harrison Schmitt, a geologist and astronaut, is
of the opinion that Helium-3 can prove to be a very useful
alternative for deuterium and tritium in nuclear fusion reactors
for power generation, as the reaction involving Helium-3 does
not result in the emission of neutrons, making it clean and nonpolluting.
Since there are not enough supplies on the Earth to provide
the Helium-3 needed to carry out commercial power generation
through fusion, in his book Return to the Moon Harrison
Schmitt argues that the moon should be mined for Helium-3.
Since the moon is continuously bombarded with solar wind, it
has a large amount of Helium-3, though in small
concentrations of about 10-20 parts per billion. Schmitt
outlines a plan to for mining helium-3 on the Moon to feed a
commercial industry for generating power, though it will be
very expensive.
Even if many consider Schmitts proposal to be unrealistic,
it is a fact that he is not the only one who thinks mining the
moon for Helium-3 is a good idea; The leader of Beijing's
space program, has said generating power via nuclear fusion
using He-3 could "solve energy demand for 10,000 years at
least". So China is seriously considering it.
B. Rare Earth Elements
Rare Earth Elements, or REE for short, are 17 elements in
the periodic table which are widely used in electronics. Their
name tends to be misleading; they are relatively plentiful in the
earths crust. [11]
Proceedings
The mining of REE was once dominated by the United

States from 1965 to 1985, until The Peoples Republic of
China started dominating the market. In 2010, a Rare Earths
Trade Dispute occurred, with China on one side and many
other countries, especially the US, on the other [10]. China had
increased the restrictions on its exports of REE by reducing its
export quota by 40% [2][12]. This resulted in a major increase
in the prices of REE outside the country. Although the Chinese
government argued that the reduction in the quota was for the
benefit of the environment (REE mining is highly polluting),
critics were unconvinced. Due to the importance of the REE in
the military, the question arose in the US of supply
vulnerability of these elements [2][13]. There were also
accusations that China had banned the exports of REE to Japan
after the 2010 Senkaku boat collision incident. The matter of
the Chinese restrictions was brought to the Dispute Settlement
Body of the WTO in 2012, by the United States. China had
joined the WTO in 11 December 2001. When it joined the
WTO, China signed an accession treaty, which did not allow
such restrictions on exports unless the goods in question were
stated, which was not the case here. China defended itself by
arguing that article XX (b) of the GATT allowed exceptions if
the restrictions were put in place for the protection of human,
animal or plant life or health. The majority of the panel was of
the view that the exceptions in article XX (b) did not justify
Chinas reduction of export quotas and that these measures
were not necessary for the protection of human, animal or plant
life or health in this case. So the WTO ruled against China in
2014 and the country removed the restrictions in 2015 [10].
There are rocks rich in Potassium, REE, and Phosphorous
along with other elements, referred to as KREEP, on the
moons surface. According to the deputy director of the Rock
Mechanics and Explosives Research Centre at the Missouri
University of Science and Technology in Rolla, Leslie
Gertsch, although REE themselves arent detectable on the
moons surface, KREEP is. Detection of components of
KREEP, such as Thorium, on the lunar surface helps locate
associated REE due to similar geochemical properties that
caused them to crystallize under the same conditions, Gertsch
says. Dale Boucher, director of innovation at the Canada-based
Northern Centre for Advanced Technology Inc., in Sudbury,
Ontario, stated that the feasibility of mining REE on the moon
can only really be judged after exploration intended for this
exact purpose. It is difficult to assess it otherwise, as there are
many factors which determine its feasibility, including cost of
separating REE obtained from ores from each other, cost of
transportation of said REE, cost of refining them etc. (North
America seems to have abundant REE deposits but it is just not
profitable to mine them). But it is still in the realm of
possibilities. If a country does indeed intend to mine REE from
the moons surface in the near future, theres a risk that a
scenario similar to the Rare Earths Trade Dispute might occur
C. Colonization
The United States, Japan, and Russia have shown interest in
the establishment of colonies on the moon [16]. The plans
seemed to have gained more ground on November 13, 2009,
137
when NASA announced the discovery of ice caps at the south

pole of the moon [15].
The reasons given by NASA for the exploration of the moon
are:
1. Human Civilization: Extend human presence to the
Moon to enable eventual settlement
2. Scientific Knowledge: Pursue scientific activities that
address fundamental questions about the history of
Earth, the solar system and the universe; and
therefore, about our place in them
3. Exploration Preparation: Test technologies, systems,
flight operations and exploration techniques to reduce
the risks and increase the productivity of future
missions to Mars and beyond
4. Global Partnerships: Provide a challenging, shared
and peaceful activity that unites nations in pursuit of
common objectives
5. Economic Expansion: Expand Earth's economic
sphere, and conduct lunar activities with benefits to
life on the home planet
6.
Proceedings
D. Future Missions of Major Parties:

[20]
1. United States (NASA)
a. Exploration Mission-1: An unmanned flight
of the Orion (MPCV), using NASAs Space
Launch System (SLS). The aim will be to
qualify the Heavy Lift Launch Vehicle (HLV)
and Beyond Earth Orbit (BEO) Orion for
transportation of humans into deep space. The
estimated date of launch is December 17,
2017. The mission is expected to last 7-10
days [18].
b. Exploration Mission-2: The first manned
mission of NASAs BEO. The crew will
consist of 4 people. The aim will be to test
Orions life support systems and to validate
its crew operations. The mission is expected
to be carried out by the year 2021 [19].
2.
Public Engagement: Use a lively space exploration

program to engage the public, encourage students and
help develop the high-technology workforce that will
be required to address the challenges of tomorrow
[3][4][5]
The 1st and 3rd reasons show their intention to colonize the
moon in the future. In January 2012 Republican candidate for
President of the United States of America, Newt Gingrich,
proposed the establishment of a US moon colony by 2020,
though the plan faced much criticism.[16]
Russias intentions seem to be different regarding the 3rd
reason. The Moon is not an intermediate point in the race; this
process has the beginning, but has no end. We are going to the
moon forever. stated Russias Deputy PM Dmitry Rogozin
who is in charge of space and defense industries. In 2007, the
Soviet Union announced its plan to establish a permanent
moon base by 2025 [14]. The main focus seems to be lunar
tourism [16]. According to a government draft that Izvestia, a
Russian newspaper, claims to have obtained, Russia has a 3step plan for manning the moon [14][17]:
1. Sending a robotic craft to the moon, possibly as early
as 2016,
2. Sending manned missions to orbit the moon by 2028,
3. A base will be set up by 2030, using resources from
the moon.
3.
a.
b.
c.
4.
Change 4: A backup probe for Change 3

[6], which put which put a lander and the
Jade Rabbit rover on the moon in 2013. The
expected year of launch is 2020. [26][28]
Change 5: To return lunar samples [27][29]
Change 6: A backup for Change 5. [27]
Japan (JAXA)
An attempt at Japans first lunar landing, an
unmanned flight, is planned for 2019, though the
budget for the plan is not available yet. [30][31][32]
5.
Japan announced its plans of setting up a moon base in
2006. They estimated that the base in question will be
established by 2030. Their main motivation, they claim, is the
development of robotics. [9]
Russia (ROSCOSMOS)
a. Luna-25: To land on the South Pole of the
moon, and test the lander technology and
communications systems
b. Luna-26: To orbit the moon for detailed
mapping of the moons surface and search for
a new landing site for future missions and to
measure the lunar atmosphere.
c. Luna-27: To go to the south pole of the moon
for testing drilling systems and analyzing
contents.
d. Luna-28: To return soil samples to the earth
e. Luna-29: To use a rover to take samples from
different locations on the moons surface
[21][22][23][24][25]
China
India (ISRO)
Chandrayaan-2: Terrain mapping by the employment
of new technology for better imaging of the lunar
surface.
138
6.
[33][34]
European Union (ESA)
Composition: metallic iron mixed

magnesium-silicates.
Found in: the main belts inner region.
ESA plans to send a lander near the moons South

Pole in 2018. The purpose of the mission is to probe
the moons surface and to test new technologies for
future landings.[7]
7.
Proceedings
with iron- and
3. M-type (metallic):
Albedo: 0.10-0.18
Composition: mainly metallic iron.
Found in: the main belts middle region.
[35]
Private
A. Motivations for Asteroid Mining

Based on consumption rates in both developed and
developing countries, it is estimated that terrestrial reserves of
elements needed for modern industry will be exhausted within
5060 years. Consequently, 44 elements are expected to face
supply limitations in the coming years, including phosphorus,
antimony, zinc, tin, silver, lead, indium, gold, and copper. All
these elements are crucial for industry [36][37][38]. Some of
the applications of a few of these elements are noted in the
following table:
Google Lunar X Prize: Astrobotic Technology, a

privately held American company, plans on launching
a private launcher in 2015, with the aim of winning the
Google Lunar X Prize, which was created in 2017 to
encourage space entrepreneurs to create a new era of
affordable access to the Moon and beyond. [8]
E. Conclusion
As it can be vividly seen, there are several detailed and
ambitious projects ready to explore the moon and benefit from
its resources. Keeping that in mind it becomes more important
than ever to firmly establish policies right now that are
acceptable to all parties to avoid any conflicts in the future.
TABLE 2.ELEMENTS
ASTEROID MINING
Besides the moon, asteroids also have potential resources
which scientists are looking to explore.
Most asteroids in the solar system are found between Mars
and Jupiter. This region is termed the asteroid belt or main belt.
Asteroids with orbits that bring them within 1.3 AU (121
million miles/195 million kilometres) of the Sun are known as
Earth-approaching or near-Earth asteroids (NEAs). These are
the asteroids which people are looking to mine. Most or all
asteroids found in the main belt appear to be NEAs.
Asteroids are usually categorized according to their albedo
(how reflective they are), composition derived from spectral
features in their reflected sunlight, and inferred similarities to
known meteorite types.
FROM ASTEROIDS AND
MAJOR APPLICATIONS
Element
Applications
Copper
. Used in electrical equipment

such as wiring and motors.
. Copper Sulfate is used as an
agricultural poison and as an
algaecide in water purification.
. Fehlings solution, a solution
containing copper, and other
copper containing solutions are
used in chemistry for qualitative
analysis.
. Copper alloys are used in
making coins.
Zinc
. Used to galvanize objects to

prevent rusting.
. Used to produce die-castings.
. Used in alloys such as brass,
nickel silver and aluminum
solder.
. Zinc oxide is used in the
manufacture of paints, rubber,
cosmetics,
pharmaceuticals,
plastics, inks, soaps, batteries,
textiles and electrical equipment.
. Zinc sulfide is used in making
luminous paints, fluorescent lights
There are 3 main categories:

1. C-type (carbonaceous): More than 75% of known
asteroids belong to this category.
Albedo: 0.03-0.09
Composition: Thought to be similar to the suns
composition, except for the absence of hydrogen, helium and
other volatile elements and compounds. Because of their low
temperatures, they are estimated to contain up to 22% water.
Found in: the main belts outer region.
2. S-type (silicaceous): About 17% of known asteroids
belong to this category.
Albedo: 0.10-0.22
139
alone. Humans will need the practice of mining in space with

robots to learn how to innovate and to work efficiently,
reliably, and safely. [42]
and x-ray screens.

Lead
Tin
Proceedings
. Used for car batteries, pigments,

ammunition, cable sheathing,
weights for lifting, weight belts
for diving, lead crystal glass,
radiation protection and in some
solders.
. Used to store corrosive liquids.
. Used in architecture, for roofing
and in stained glass windows.
2. Planetary Resources (Private, US)

The company announced on April 24, 2012 its plans to
mine NEA. They plan to launch prospective spacecraft for
collecting data about asteroids within 7 years. The company
plans to build new technology which will reduce the price of
space travel so that it can be successful in its mission.
[46][47][48]
4.
. Used to coat other metals to

prevent corrosion.
. A niobium-tin alloy is used for
superconducting magnets.
. Tin salts sprayed onto glass are
used to produce electrically
conductive coatings.
Deep Space Industries (Private, US)
On January 22, 2013, this private company announced its

plans to launch spacecraft in 2015 for prospecting suitable
asteroids and returning samples to the earth, and to begin
mining NEA by 2023. [43][44][45]
5.
Kepler Energy and Space Engineering (KESE)
They aim to build and send a 4-module Automated Mining

System (AMS) to a small asteroid with a simple digging tool to
collect approximately 40 tons of samples by the end of the
decade, using existing guidance, navigation and anchoring
technologies.[51]
[39]
B. Mining Considerations
There are three options when considering mining asteroids:
1. Bringing the raw material to earth.
2. Processing the raw material on the asteroid itself and
then bringing the processed material to earth.
3. Transport the asteroid to a safe orbit around the
Moon, Earth or to the ISS. This method is estimated to be the
one which allows the most materials to be used.
[40][41]
C. Proposed Mining Projects
There are four main parties currently proposing projects for
mining asteroids:
1. NASA (US)
An upcoming mission, termed OSIRIS-Rex, scheduled for
launch in 2016, will involve travelling to a near-Earth asteroid
called Bennu and bringing a small (about 60g) sample back to
Earth for study. The destination will be reached in 2018 and the
sample will be transported to the earth in 2023.[49][50]
NASA Institute for Advanced Concepts (NIAC) announced
the Robotic Asteroid Prospector (RAP) project on in
September 2012. The purpose of RAP is to evaluate the
feasibility of asteroid mining in terms of means, methods, and
systems. Their stated strategy is: Initial prospecting missions
will be robotic, although early industrial mining missions will
require both humans and robots on-site. There will be too many
judgment calls and qualitative decisions on-site for robots
D. Conclusion
It becomes clear after looking at the motivations and future
plans of various parties (countries as well as private parties)
that a future marathon for space resources is highly probable.
Unfortunately, the Outer Space and Moon Treaty hinder the
steps made towards the mentioned endeavours as they face
many legal barriers. For now, only the hurdles provided by the
outer space treaty will be discussed, as the moon treaty has not
been signed by major influential parties. Specific parts of the
two treaties will be discussed in more detail later on in the
paper.
Based in part on the Antarctic Treaty, the Outer Space
Treaty only appears to permit research activities on the moon
and other celestial bodies, though this is debatable. The idea
that you can't claim sovereignty is not necessarily incompatible
with the right to go conduct mining operations. The high seas
are not subject to any sovereignty, but people can go and fish
there. said international lawyer and space-law expert Timothy
Nelson in an interview with space.com, referring to an article
in the 1967 treaty stating: Outer space, including the moon
and other celestial bodies, is not subject to national
appropriation by claim of sovereignty.
Even then, the treaty hinders research, mainly owing to a
lack of incentive. Due to the statement that no particular party
can claim any ownership rights to the resources, there will be a
lack of commercial concern, ultimately leading to a lack of
funding for research, hence lowering the amount of potential
research which can be carried out.
140
HISTORICAL ANALYSIS
The ever-evolving economic landscape presents a unique
challenge in the world of politics. The one of keeping peace
and making sure that the resources of a state are not
illegitimately taken away from it. The solution is an old one;
treaties and organizations to enforce the treaties.
In this section conflicts and resources as a main reason for
conflicts will be discussed, along with how they can be
avoided. This will lead to the importance of treaties in the
international landscape and through the analysis of treaties and
treaty organizations a parallel will be drawn between the
historic treaties and conflict and the current space situation to
show that the current treaties governing space exploration and
resources management are not viable.
A. Wars
History has seen many wars fought over resources. Hitlers
campaign in 20th century, before the Second World War is a
perfect example of a war fought over resources. When Hitler
came to power in 1933 his whole agenda centered on making
Germany a powerful sovereign country again. There were two
important aspects to this. One was regaining areas with great
economic potential, such as Alsace and Loraine on the western
border; they were highly industrialized zones and had Iron ore
and Coal deposits. Another was to get more land. This was a
part of Lebensraum (literally living space), a German
ideology proposing expansion to give more space to the
German people. This generally advocated expansion on the
Eastern front. [52]
Another event is the period in history known as the
Scramble for Africa between 1881 and 1914. It is the period
of invasion, occupation, colonization and annexation of
African territories by European empires. Africa had a lot of
natural resources, ranging from coal to diamonds, and a huge
untapped potential. Furthermore the African population was
also used as a source of free labor through slavery. As
expected, with such resources available, it was not long before
war broke out, not only the war against the natives to take
control of the area, but also between different European powers
to get the biggest piece of the pie. [53]
The most recent, and one that is still relevant today, is of
the Oil War. Oil war is the term given to all the conflicts over
petroleum resources. Oil has become very important in the
modern world, as primary source of energy its significance is
unprecedented. Petroleum resources alone can run a countrys
economy such as the Middle East states, while on the other
hand import of oil can cripple a country. The result of all this is
that oil became an extremely important resource in the
international landscape, a resource that countries wage wars
over. There have been a number of conflicts over oil [54]. The
Iran-Iraq war (1980-1988), the Gulf War (1990-1991) and the
Afghan War (1978-present) are major examples of this. Some
conflicts that have control of petroleum resources as an ulterior
Proceedings
motive are still present today. The most recent major crisis was
Russias annexation of Ukraine, this not only demonstrates
Russias interest in controlling the oil pipeline routes, but also
that of United States [55]. And perhaps the most important is
the creation of ISIS. ISIS sits on an oil rich land, and even
though no state can legally buy oil from them, they do have a
steady income through oil sales. This location of ISIS and the
possible effect of their oil sales made it important to other
foreign powers even before they began global terrorism.
[56][57].
As it can be seen, resources are an impetus for war. And as
the previous section elaborated, space has a lot of resources
available and a space race between various parties is highly
probable. Whereas history shows, there is an equally high
probability of conflict as a result of this. And in order to avoid
this we need a comprehensive treaty system.
B. A Failed Treaty
As with anything else, it is important to learn from the
mistakes already made by others. The same goes with treaties.
It is crucial to see where the previous treaties went wrong and
to rectify them. In hope of avoiding future conflicts from
turning violent and it is of immense importance when
considered that now the domain of humans is not just restricted
to the Earth, but also includes the space.
The Treaty of Versailles of 1919 is one of the most
important treaties ever signed. It effectively brought an end of
the First World War, though unknown at the time it was
signed, it was a significant factor in the start of the Second
World War. [58]
The treatys ultimate failure was when Germany invaded
Poland in 1939, dragging the whole of Europe into the Second
World War. There were a number of reasons for the failure of
the treaty. One was that the treaty had been made by the
winners of the war and it was very profitable for them i.e. it
was a biased treaty where everyone was not on the table. The
treaty was generally against one country (Germany) and the
actors at the time felt that as long as one rogue state was
controlled, the peace would not be disturbed, but in reality
limitations were needed for all the countries. Another reason
was that there was no way to reinforce the terms of the treaty,
such as the clauses that called for all-around disarmament.
None of the countries wanted to reduce armaments, so no one
was there to enforce the decision either. And whereas countries
like Britain and France had agreed to keep Germany in check,
they were not ready to go to war for it or even impose sections
that would hurt their own economy. Hence it can be seen that
the ultimate weakness of the treaty was that there was no one to
stop any country from breaking it.[52]
Most of these reasons are generic, and very similar reasons
lie behind the failure of almost all treaties. And so when
planning for the future, especially in the case of forming
141
treaties that have universal importance, like as those related to

the space, such weaknesses in treaties of the past should be
avoided. A few of the common shortcomings that must be
avoided are the treaty should not be directed against just a
certain country or to control just the rogue states, but be
equally applied to everyone; furthermore the treaty needs to
absolutely binding, and the powerful states that promise to
direct it, must be legally compelled to do. And most
importantly there should be measures that make it impossible
for states to violate or ignore the treaty.
the Antarctica Treaty System immensely important. Signed in

1959, the treaty established a few ground rules for every state
regarding Antarctica. First off the geographical area of
Antarctica was determined. And it was agreed that the area will
be used for peaceful purposes only. Freedom of scientific
investigation and cooperation was guaranteed. And importantly
it prohibited new territorial sovereignty claims, but did not
dispute nor recognize the old claims. [60]
C. A Failed Organization
An organization created along with Treaty of Versailles
was the League of Nations. It was the predecessor to the United
Nations. It had two main aims, one to maintain international
peace through collective security and the second to promote
international cooperation to solve social and economic
problems. [59] The former was the more important function of
the body. The world had just come out of the First World War,
and sought peace and stability. The main idea was to discuss
the problems as they rose, and solve them through an
understanding of all the countries through a unanimous effort.
Though the League looked very good on paper, it was not so
successful in reality, with the Second World War an epitome of
its failure. There are a number of reasons why the organization
failed. Firstly it did not include all the countries, major
exclusions were the US, Germany, and USSR. And there were
serious weaknesses in the covenant, like unanimous decisions
were needed and the League had no force of its own. The result
was that it could advise on how to solve problems, but it could
not force a state. Furthermore it was a very British and French
affair; these were the two most powerful countries at the time
and the league was always under their control. As a result the
League was prone to letting allies of the major states get away
with anything; major examples of this are the Japanese
invasion of Manchuria (1931) and the Italian invasion of
Abyssinia (1935). [52]
Once again it is important to take care of these few
parameters in order to form a comprehensive organization that
would be much more equipped to deal with problems in a
manner that would avoid conflict. The Regime is an
organization was proposed by the Moon Treaty to for the use
of the Moon and other celestial bodies, and it is yet to be
formed. Keeping the weaknesses of past organizations in mind,
especially those that led to the failure of the League of Nations,
the framework of the Regime should be such that it covers all
states equally and does not hinder it in its function, yet allows
it to be decisive. All countries should get a say in it, but
decisions should not be unanimous as it would render it
impractical, and similarly it should not give powerful states the
veto power, as it too could hinder the organization.
D. A Successful Treaty
Antarctica is the closest example of space on Earth.
Antarctica, like space, has no native inhabitants. This makes
Proceedings
Further additions were made to the initial treaty, among

them, the most important was the Convention on the
Regulation of Antarctic Mineral Resource Activities (1988),
which made mining illegal in Antarctica even though there coal
deposits and hydrocarbons along with other minerals [61]. And
the Protocol on Environmental Protection to the Antarctic
Treaty (1991) was to insure the protection of the Antarctic
environment and ecosystem.[62]
The Antarctic treaty has been a success so far. It has
managed to peacefully promote Antarctica as a research area
and kept it from being militarized. And it has also successfully
prevented any conflicts over territory by not recognizing or
disputing preexisting claims, and forbidding new claims.
Furthermore the environment of Antarctica has been protected
and it remains the only continent that has not been affected by
the human population.
The Antarctic Treaty has been successful, however that
does not make it a perfect model to work with in terms of a
space treaty. Unlike Antarctica, outer space is not a small
potential contributor to worlds demands for mineral resources.
Moreover, outer space also does not pose any environment
threats to Earth like industrialization of Antarctica could.
Furthermore, research in space might very well soon depend on
extraction and utilization of space resources to fuel it. Yet on
the other hand, the Antarctic Treaty has prevented conflict in
the region in an exemplary way, and the features of the treaty
that guaranteed this should be replicated which will be
discussed latter.
The current treaties are inspired by the Antarctic Treaty but
success cannot be replicated this way since the conditions are
far too different. As a result, the UN must recognize the
difference to move forward to full utilization of outer spaces
potential, yet in such a manner that global peace is not
jeopardized.
FLAWS IN THE TREATIES
A. The Res Communis Controversy
One of the chief controversies is the Res Communis
ideology employed especially in the Moon Treaty 1979 [65].
The phrase common heritage of mankind contributes to the
overall vagueness of the treaty and hence fuels the insecurity of
interested nations. Since moon and celestial objects are a
province of mankind [66], any form of active military
involvement is forbidden. However, considering the rise of
terrorist organizations like ISIS and rogue nations like North
142
Korea, both increasingly using sophisticated technology; it

seems dubious that space would remain untainted from terrorist
activities in the future.
In the case of USA, however, the reasons for disagreement
were political. The treaty put developed countries in a place
where they or their private companies could be attacked for
using outer space resources by developing countries. This
concerned scientists, entrepreneurs and politicians alike as it
not only meant a decreased interest in outer space by the
government and private groups but also a resulting decline of
space science research because if the moon and celestial bodies
are to remain unusable, then there is little incentive for them to
work.
It is true, however, that it could be done under the watch of
an international regime [67] but that too raises the fear of
blackmail since non-space-faring countries are in a majority
and most likely to lobby themselves to obtain economic
advantages. It is expected that the international regime would
be similar to The Enterprise proposed in Agreement of the
Law of the Sea Convention 1994[68]. The Enterprise, by its
very nature, was to receive a portion of wealth obtained by the
parties exploiting marine resources and distribute it amongst
developing countries. Such a system could prove discouraging
for space resources as there is a much higher capital required
and much greater risk of failure, combining this with low
profits means that space resources may remain untouched for a
long time even after sufficient technology has been developed.
Outer Space Treaty enforced the idea that no state could
claim any region in space [69], however it was not very
specific about private companies and groups. But the Moon
Treaty attempted to make sure that even private groups could
not lay claim to any region in space. As expected, it raised a lot
of opposition from the business community and it became one
of the most controversial points of the treaty.
Proceedings
developed a significant space involvement and hence deem the

development of such policies to be pointless in the present, as
the delegate of Argentina replied to COPUOS question
regarding definition of Outer Space: The present magnitude of
space and aviation activities and technologies has not given
rise to a need for the adoption of specific provisions delimiting
outer space [70].
Though Argentina has CONAE as its space agency, its
future plans, however, do not include any form of space
exploration, simply satellite management [71], unlike countries
like China, Japan, USA, Russia and India as elaborated in
Section II. The case of Argentina is not singular but it
represents the hesitation of many countries that have too weak
space involvement to be able to make decisions concerning
their role in it. Hence to play it safe they decide not to indulge
in the process of any policy making which they may find even
slightly restrictive in the future.
Consequently, a large number of developing and some
developed countries are not a part of policy making because
their national experts and politicians do not have the
knowledge and understanding of the importance and
significance of space policy making, especially those
pertaining to resources because since Moon and celestial
bodies have been declared a common heritage of mankind it is
important that all countries understand their resulting rights and
limitations
B. Property and Ownership

It has been agreed that ownership of space and celestial
objects is not possible for countries but there is an apparently
silent approval for private groups. This raises further questions
and controversies.
Firstly, how can the ownership be declared if there is no
national government to maintain and uphold the rights of the
property owner. Such a situation would force the owners to
establish their own system of defense which may lead to an
arms race or recurring physical conflict in space.
As it is later suggested, the establishment of an organization
like The Enterprise is critical. Moreover, the UN should be
prepared to accept the eventual necessity of a peace-keeping
force in space to assist the organization in enforcement of its
mandate.
D. Confidentiality and Privacy

The Moon Treaty states all space vehicles, equipment,
facilities, stations and installations on the moon shall be open
to other States Parties [72]This could give rise to a lot of
possible misuses and controversies such as technology theft,
infringement of privacy and right to confidentiality.
If it is possible for any state to inspect the equipment of any
other state then it is possible that this may be used to replicate
the technology of the first state. It poses many questions about
the right of having intellectual properties in space. Even though
countries and private groups may accept space as a common
heritage but it will be unfair for them to lose the rights to all
their individual research and development in space. Since such
research is extremely expensive, it is very likely that such a
jurisdiction would very severely dent the interest in it.
Parties operating in space and celestial bodies will also
have to disclose the specifics of their activities even though
they may be well within their rights. It too opens up debate on
human rights, among other things. If, for example, a researcher
is studying the health of astronauts on moon, he may have to
disclose data regarding this research, which may contain
personal information of astronauts. This can be taken as a
violation of Universal Declaration of Human Rights [73].
C. Lack of Space Involvement

On the other end of the spectrum, another hurdle in
development of sustainable space policies is a lack of
awareness and interest by some countries stemming from a
lack of space involvement. These countries often have not
E. Lack of Violation Protocols

Both the treaties fail to outline a mechanism that is to be
followed if the treaties are breached by a country. United
Nations would most likely be unable to undertake any legal or
physical action since the Security Council would need
143
unanimous votes to do so. As a result the treaties appear to

restrict less influential and less powerful countries only, since
the stronger ones can easily get off without any rebuttal as they
can diplomatically procure a veto.
The example of Antarctic Treaty needs to be considered
here [74]. Interestingly enough, Antarctic treaty was the
inspiration for Outer Space treaty and yet such an integral
component was left out. According to Antarctic Treaty, in case
there is a clash of purposes between two parties, the matter can
be taken into International Court of Justice from where the
matter could be resolved. Not only does it safeguard the
integrity of the treaty but it also makes it more trustable, one of
the many reasons of its success.
The two treaties were missing a core element of a
successful international agreement, the violation protocols and
as it was witnessed with the failure of Moon Treaty, other
articles are meaningless without them.
F. Stifling Space Research and Funding
It is true that Outer Space Treaty intends to maximize
research and interest in outer space; however, its actual effects
have been different. As it has been previously established, the
restrictions collectively work to discourage investors. This in
turn leads to a direct decrease in the funding of researchers and
scientists who are researching on space related areas. Apart
from nationally run agencies, few research groups are able to
sustain themselves. Consequently, the research that has been
done is nowhere near the research that could have been done
without the policies.
If some of the most restrictive clauses are removed, there
would be an immediate rise in research, jobs, interest and
awareness regarding space and celestial bodies. Space is an
area that captures the imagination of many young scientists but
they are often unable to find employment or even advanced
education in their area of interest.
SOLUTIONS
A. Establishment of Extraction-Friendly Framework
The current policies need to be modified to encourage both
the private and government space agencies to attempt
extraction of resources from regions beyond Earth. To that end,
several changes need to be made.
Private parties should be allowed to have at least temporary
property rights so their reservations about lack of control can
be met. It makes sense to allow property rights on, for example,
asteroids, because after the extraction is over there will be little
left of it.
Parties should also have a greater degree of privacy and
confidentiality. Specific rules and regulations should be framed
to ensure that the rights to intellectual property and personal
privacy of astronauts are conserved.
Moreover the specifics of a regime like The Enterprise
for space should be discussed and decided. Special attention
should be given in making sure that space agencies are not
Proceedings
restricted unfairly by the share they have to provide to such a

regime.
The shares from the extraction should be based on a model
such that small scale extraction is bound to a lesser percentage
and large scale extraction to a greater percentage.
B. Increasing Space Awareness
Established agencies like the NASA should work harder to
develop a sense of understanding and importance of space
resources and hence the importance of policy making
governing it.
Only after the governments and experts of individual
nations realize the significance of such policies and how they
are critical in avoiding future clashes can the UN as a whole
move forward.
It may also be important to get international public support
for the policies to pressurize the governments into making
decisions in the present than letting them hang on for the
future. This requires the use of education and awareness
mechanisms through television, newspapers and public talks.
C. International Space Property Rights
Due to the sheer enormity of space, abundance of celestial
bodies and rapid shifting, it may never be possible to divide
regions of space in a manner that is fair to all. However, the
common heritage ideology may prove extremely impractical in
the future. As an analysis of history shows, such a concept is
extremely nave when considered alongside the thousands of
wars that people have waged for land and resources.
Space agencies, both private and government, need to have
a tangible incentive for them to risk a great amount of capital
and manpower in attempting to reach for the moon and beyond.
Loss of the right to ownership can discourage space
exploration. They should be entitled to own, use and sell the
resources that they have extracted, to at least some degree.
However, concerns of developing nations should also be
addressed pertaining to property rights. Their current reduced
state of development does not mean that they lose any chance
of being a part of the future.
Taking both sides into consideration, one method that
seems feasible is to employ a procedure whereby regions of
space are provided as property only during the period they are
being actively used by the party to discourage gluttony.
.
D. Developing Violation Protocols
To uphold the good intentions, enforce articles and develop
trust, there needs to be a strong and independent system of
justice which ensures that the rights of less powerful and less
influential countries remain preserved.
To that end, a section of the International Court of Justice
could be formed to deal with space law. However, for its
smooth operation the policies need to be much more detailed.
In fact, it would be wise to draw up an internationally
acceptable constitution governing activities in space and
celestial objects. Not only would it act as a code of conduct for
144
[15] Pat Cahill. (2014, January 20). Mount Holyoke professor Darby
parties interested in space resource extraction but would also

help quickly resolve any minor conflicts that may arise.
E. Using International Diplomacy Positively
Moon Treaty was treated very coldly by the international
community. However, an active interest of a single party such
as the USA could have totally reshaped its outcome as the
decisions of such influential nations are followed by their close
allies as well. In fact, an agreement on an international issue
between USA, China and Russia almost always gains nearly
unanimous global support.
As a result bloc leaders like USA and China should take a
responsive and positive role in the development of space
policies. By considering the interests of their allies, many of
those who do not have an independent space program, they can
help shape a globally acceptable and beneficial treaty.
REFERENCES
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
Proceedings
[16]
[17]
[18]
[19]
[20]
[21]
[Online] Available: www.islandone.org/Treaties/BH766.html

Leonard David. (2010, October 4). Is Mining Rare Minerals on
the Moon Vital to National Security? [Online]. Available:
http://sservi.nasa.gov/articles/is-mining-rare-minerals-on-themoon-vital-to-national-security/
Beyond
Earth
[Online].
Available:
http://www.nasa.gov/exploration/home/why_moon.html
(December
2006).
Lunar
Exploration
Objectives
(1sted.)[Online].
Available:
http://www.nasa.gov/pdf/163560main_LunarExplorationObjecti
ves.pdf
[Online]. Available: http://www.permanent.com/lunar-miningthe-moon.html
(2013, December 16). China plans to launch Chang'e 5 in 2017
[Online].
Available:
http://news.xinhuanet.com/english/china/201312/16/c_132971252.htm
(2010, September 16). NEXT STEP FOR ESAS FIRST
MOON
LANDER
[Online].
Available:
http://www.esa.int/Our_Activities/Human_Spaceflight/HSF_Dir
ectors_Corner/Next_step_for_ESA_s_first_Moon_lander
[Online]. Available: http://lunar.xprize.org/
(2003, August 3). Japan Plans Moon Base By 2030 [Online].
Available:
http://www.moondaily.com/reports/Japan_Plans_Moon_Base_B
y_2030_999.html
China Measures Related to the Exportation of Rare Earths,
Tungsten
and
Molybdenum
[Online].
Available:
https://www.wto.org/english/tratop_e/dispu_e/cases_e/ds431_e.
htm
Kieran Mulvaney. (2012, March 14). The Fight Over Rare
Earths:
A
Primer
[Online].
Available:
http://news.discovery.com/earth/the-fight-over-rare-earths-aprimer-120314.htm
Lesley Stahl. (2015, March 22). Modern lifes devices under
chinas
grip?[Online].
Available:
http://www.cbsnews.com/news/rare-earth-elements-chinamonopoly-60-minutes-lesley-stahl/
Leonard David. (2010, October 4). Is Mining Rare Minerals on
the Moon Vital to National Security? [Online]. Available:
http://www.space.com/9250-mining-rare-minerals-moon-vitalnational-security.html
(2014, May 9). Russia will begin Moon colonization in 2030 draft
space
program
[Online].
Available:
http://rt.com/news/157800-russia-moon-colonization-plan/
[22]
[23]
[24]
[25]
[26]
[27]
[28]
[29]
[30]
[31]
[32]
145
Dyar says a colony on the moon is no fantasy [Online].

Available:
http://www.masslive.com/living/index.ssf/2014/01/mount_holyo
ke_professor_part_of_massive_study_of_outer_space.html
DeAnnaKerley. 15 Ambitious Plans to Colonize the Moon
[Online]. Available: http://mentalfloss.com/article/53588/15ambitious-plans-colonize-moon
Anna Dolgov. (2014, May 8). Russia Plans to Colonize Moon
by 2030, Newspaper Reports
[Online]. Available:
http://www.themoscowtimes.com/article/499806.html
Chris Bergin. (2012, February 29). Exploration Mission 1: SLS
and Orion mission to the Moon outlined [Online]. Available:
http://www.nasaspaceflight.com/2012/02/exploration-mission-1sls-orion-debut-mission-moon-outlined/
Chris Bergin. (2012, October 30). EM-2: Orion crew to spend
up to four days in Lunar Orbit [Online]. Available:
http://www.nasaspaceflight.com/2012/10/em-2-orion-crewspend-four-days-lunar-orbit/
Future
planetary
missions
[Online].
Available:
http://spaceexp.tumblr.com/future-missions
Russian Moon missions face three-year delay [Online].
http://www.russianspaceweb.com/spacecraft_planetary_2014.ht
ml
(2014, October 10). SRI: Russia will begin launching automatic
stations to the Moon in 2018 [Online]. Available:
http://ria.ru/space/20141003/1026822990.html#ixzz3eXhKtcXA
http://ria.ru/space/20141003/1026822990.html
Russian Lunar Polar Probes, 2015-2020 [Online]. Available:
http://www.permanent.com/russian-lunar-polar-probes.html
(2014, August 3) Manned Moon Mission to Cost Russia $2.8
Bln Space Research Institute [Online]. Available:
http://sputniknews.com/russia/20140803/191677434.html#ixzz3
eXiSLeokhttp://sputniknews.com/russia/20140803/191677434.h
tml
(2014, October 10) Russian Luna-25 Mission to Cost Billions:
Space
Agency
[Online].
Available:
http://sputniknews.com/russia/20141010/193919188.html#ixzz3
eXj6Nr00http://sputniknews.com/russia/20141010/193919188.h
tml
(2015, May 22). China Plans First Ever Landing On The Lunar
Far
Side
[Online].
Available:
http://www.spacedaily.com/reports/China_Plans_First_Ever_La
nding_on_the_Dark_Side_of_the_Moon_999.html
Leonard David. (2015, March 17). China Outlines New Rockets,
Space Station and Moon Plans [Online]. Available:
http://www.space.com/28809-china-rocket-family-moonplans.html
Alexandra Witze. (2014, March 19). China's Moon rover awake
but
immobile
[Online].
Available:
http://www.nature.com/news/china-s-moon-rover-awake-butimmobile-1.14906
Bradley Perret. (2013, December 16). China Plans 2017 Lunar
Sample-Return
Mission
[Online].
Available:
http://aviationweek.com/space/china-plans-2017-lunar-samplereturn-mission
(2015, April 19). JAXA plans Japans first moon shot in 2018
[Online].
Available:
http://www.japantimes.co.jp/news/2015/04/19/national/jaxaplans-japans-first-moon-shot-in-2018/#.VY5YSruqqkp
Dianne Depra. (2015, April 4). Japan Aiming For Moon
Landing By 2018 According To JAXA [Online]. Available:
http://www.techtimes.com/articles/48266/20150424/japanaiming-for-moon-landing-by-2018-according-to-jaxa.htm
(2015, April 22). JAPAN TO ATTEMPT UNMANNED
LUNAR LANDING IN 2018 [Online]. Available:
http://www.spaceflightinsider.com/organizations/jaxa/japansjaxa-to-attempt-lunar-landing-in-2018/
[33] (2010, Sep 5). Were launching Chandrayaan-2 for a total
[34]
[35]
[36]
[37]
[38]
[39]
[40]
[41]
[42]
[43]
[44]
[45]
[46]
[47]
[48]
[49]
[50]
[51]
coverage
of
the
moon
[Online].
Available:
http://timesofindia.indiatimes.com/home/opinion/interviews/We
re-launching-Chandrayaan-2-for-a-total-coverage-of-themoon/articleshow/6501413.cms?referral=PM#
India to launch Chandrayaan-2 by 2017 [Online]. Available:
http://www.thehindu.com/sci-tech/science/india-to-launchchandrayaanii-by-2017/article5562361.ece?ref=sliderNews
Asteroids
[Online].
Available:
http://nssdc.gsfc.nasa.gov/planetary/text/asteroids.txt
Stephen Shaw. (2012, August 21). Posts Tagged C-type
asteroids
Asteroid
Mining
[Online].
Available:
http://www.astronomysource.com/tag/c-type-asteroids/
Endangered
Elements
[Online].
Available:
http://www.acs.org/content/acs/en/greenchemistry/researchinnovation/research-topics/endangered-elements.html
David Cohen. (2007, May 23). Earth's natural wealth: an audit
[Online].
Available:
http://www.sciencearchive.org.au/nova/newscientist/027ns_005.
htm
Periodic Table [Online]. Available: http://www.rsc.org/periodictable/
Stephen Harris. (2013, April 8). Your questions answered:
asteroid
mining
[Online].
Available:
http://www.theengineer.co.uk/aerospace/in-depth/yourquestions-answered-asteroid-mining/1015966.article
(2012, April 2). Asteroid Retrieval Feasibility Study [Online].
Available:
http://kiss.caltech.edu/study/asteroid/asteroid_final_report.pdf
Marc Cohen. Robotic Asteroid Prospector (RAP) Staged from
L-1: Start of the Deep Space Economy [Online]. Available:
http://www.nasa.gov/offices/oct/early_stage_innovation/niac/20
12_phase_I_fellows_cohen.html
Mike wall. (2013, January 22). Asteroid-Mining Project Aims
for
Deep-Space
Colonies
[Online].
Available:
http://www.space.com/19368-asteroid-mining-deep-spaceindustries.html
(2013, January 22). Commercial Asteroid Hunters announce
plans for new Robotic Exploration Fleet [Online]. Available:
http://www.prweb.com/releases/2013/1/prweb10346181.htm
Taylor Soper. (2013, January 22). Deep Space Industries
entering asteroid-mining world, creates competition for
Planetary
Resources
[Online].
Available:
http://www.geekwire.com/2013/deep-space-industries-enteringasteroidmining-world-creates-competition-planetary-resources/
http://www.planetaryresources.com/technology/#space-basedobservation
(2012, April 24). Company aims to mine resource-rich asteroids
[Online].
Available:
http://lightyears.blogs.cnn.com/2012/04/24/companies-plan-tomine-precious-metals-in-space/?hpt=hp_t3
(2012, April 24). Plans for asteroid mining emerge [Online].
Available:
http://www.bbc.com/news/science-environment17827347
(2013, August 8). New NASA Mission to Help Us Learn How
to
Mine
Asteroids
[Online].
Available:
https://www.nasa.gov/content/goddard/new-nasa-mission-tohelp-us-learn-how-to-mine-asteroids
(2015, June 22). OSIRIS-REx Team Prepares for Next Step in
NASAs Asteroid Sample Return Mission [Online]. Available:
https://www.nasa.gov/feature/goddard/osiris-rex-team-preparesfor-next-step-in-nasa-s-asteroid-sample-return-mission
Michael Buet and Robert Frantz. ASTEROID MINING: WHY
AND
HOW?
[Online].
Available:
http://www.friendsofbest.org/wp-
[52]
[53]
[54]
[55]
[56]
[57]
[58]
[59]
[60]
[61]
[62]
[63]
[64]
[65]
[66]
[67]
[68]
[69]
[70]
[71]
[72]
[73]
[74]
146
Proceedings
content/uploads/2015/05/ASTEROID-MINING_WHY-ANDHOW-Article_03.pdf
Norman Lowe, Mastering Modern World History, 4th edition,
New York: Palgave Macmillan, 2007
Andrina Stiles, "Bismark's Germany" in The Unification f
Germany, 1st edition, London, England: Hodder& Stoughton,
1998, pp. 85-99
John Lowe and Robert Pearce, "Colonial Rivalries, 1870-1914"
in Rivalry and Accord - International Relations 1870-1914, 1st
edition, London, England: Hodder& Stoughton, 2007, pp. 39-69
Michael Klare, The Race for What's Left: The Global Scramble
for the World's Last Resources, 1st edition, New York: Picador,
2013.
Dr. Nafeez Ahmed. (2014, March 6), Ukraine crisis is about
Great Power oil, gas pipeline rivalry [Online]. Available:
http://www.theguardian.com/environment/earthinsight/2014/mar/06/ukraine-crisis-great-power-oil-gas-rivalspipelines
Keith Johnson. (2014, September 24), U.S. Strikes ISIS Oil
Installations
[Online].
Available:
http://foreignpolicy.com/2014/09/24/u-s-strikes-isis-oilinstallations/
Treaty of Versailles, Paris, 28 June 1919, League of Nations
Treaty
Series,
available
from
https://en.wikisource.org/wiki/Treaty_of_Versailles
Covanent
of
League
of
Nations,
available
at
https://en.wikisource.org/wiki/Covenant_of_the_League_of_Nat
ions
Antarctic Treaty, Washington, 1 December 1959, available from
http://disarmament.un.org/treaties/t/antarctic/text
Environmental Treaties and Resource Indicators (ENTRI),
Available
From:http://sedac.ciesin.columbia.edu/entri/texts/acrc/cramra.txt
.html
Protocol on Environmental Protection to the Antarctic Treaty,
1991,
available
from
http://www.antarctica.ac.uk/about_antarctica/geopolitical/treaty/
update_1991.php
Outer Space Treaty, Washington DC, 10 October 1967,
available
from
http://www.unoosa.org/oosa/en/ourwork/spacelaw/principles/leg
al-principles.html
Moon Treaty, New York, 18 December 1979, available from
http://www.unoosa.org/oosa/en/ourwork/spacelaw/treaties/moon
-agreement.html
Para 1, Article 11, Agreement Governing the Activities of States
on the Moon and Other Celestial Bodies, United Nations, 1979
Article 1, Treaty On Principles Governing The Activities Of
States In The Exploration And Use Of Outer Space, Including
The Moon And Other Celestial Bodies, United Nations, 1967
Para 5 -7, Article 11, Agreement Governing the Activities of
States on the Moon and Other Celestial Bodies, United Nations,
1979
Part XI, Agreement of the Law of the Sea Convention, United
Nations, 1994
Article 2, Treaty On Principles Governing The Activities Of
States In The Exploration And Use Of Outer Space, Including
The Moon And Other Celestial Bodies, United Nations, 1967
A/AC.105/889/Add.14, Committee on the Peaceful Uses of
Outer Space, United Nations General Assembly
CONAE website : http://www.conae.gov.ar
Para 1, Article 15, Agreement Governing the Activities of States
on the Moon and Other Celestial Bodies, United Nations, 1967
Article 12, Universal Declaration of Human Rights,1948
Para 2 Article 11, The Antarctic Treaty, 1959
Proceedings
Feasibility Study to Install Firefighting Equipment on

a Cargo Helicopter
S. Shahid1, B. Saeed2, S. Javed3
Department of Aerospace Engineering
Pakistan
1
msohailshahid135@gmail.com
3
saad.javaid@gmail.com
Abstract Aerial firefighting has seen recent interest with
events such as forest and domestic fires seen on the rise
across the world. Resources such as fixed wing aircraft and
helicopters have been employed for use in firefighting
setups. Aerial firefighting began around 1920 with first
attempts at dropping water aircraft onto a fire. Most of
these attempts were unsuccessful during this era, but they
provided a platform for future initiatives. In 1935, the
Aerial Fire Control experimental Project was created. At
this point, aircraft became important for fire detection.
Inspired by agricultural spraying techniques, helitanker
soon became the next progression in aerial firefighting.
With firefighting helicopters came heli-rapellers
firefighters who rappelled down to the fire as a ground
force. In the 1960s, surplus military aircraft became
another effective way to combat wildfires. Planes like the B25, Douglas B-26 and Lockheed PSV could carry 1000
gallons of water. Similarly some helicopters use fire watch
cameras for the surveillance of fire. Its infrared thermal
imager can detect heat of a wild fire even through thick
smoke. The aim of this project is to carry out feasibility
study of installation of firefighting equipment on MI-171
helicopter and determine its practicability. A study will be
carried out that whether MI-171 can be used for firefighting
purposes. MI-171 has the main advantage of its reliability.
By virtue of its qualities MI-171 is successfully operated in
different conditions whether it is heat or cold, rain or
snow, mountings or desert, sea or land. It has wide range of
mission accomplishments like transportation of up to 4000
kg cargo inside cargo compartment, transportation of up to
4000 kg cargo on external sling, SAR missions, fire-fighting,
patrolling, construction works, offshore oil rig flights,
ambulance mission etc. Keeping all the above traits in view
MI-171 can be made capable of fighting the wild fires. As it
has the ability to carry a lot of weight inside its cargo cabin
and mounted on external sling, these areas will be analyzed
especially. For this first of all volume and weight capacity
of cargo section of MI 171 helicopters will be carried out.
After that all the forces applied on the helicopter structure
will be estimated. Based on these forces the aircraft
modified CAD model will be analyzed numerically in
ANSYS software. Main focus will be on the structural and
aerodynamic analysis of the helicopter with water tank
installed inside the cargo section. For structural analysis,
Workbench software will be used. The results will be

validated by solving the problems analytically.
Index Terms Aerospace, Firefighting, Helitanker Stress
analysis
INTRODUCTION
Wildfires are natures most dangerous, terrifying and
recurring phenomenon. Similarly periodic wildfires are a
natural part of echo system dynamics. A large fir has an average
cost of between $2.1 million and $4.5 million (which includes
suppression costs, social costs and the magnitude of insured
losses). On the other hand a successful initial attack on the same
large fire saves $3.3 million. Aerial firefighting has seen recent
interest with deadly events such as forest and domestic fires
went on the rise across the globe. Resources such as fixed wing
aircraft and helicopters have been employed for use in
firefighting setups. Aerial firefighting is the use of aerial
resources to suppress fires. Many aircrafts like fixed-wing,
helicopters, smoke jumpers and rappellers are used for this
purpose. A wide variety of terminology is used in the media for
the aircrafts used in aerial fighting. The term air tanker is used
for the fixed wing aircraft generally. One of the major
contributor towards aerial firefighting is helicopter. Helicopters
can be fitted with buckets or tanks, called helitanker for
firefighting purpose. Some of them are also fitted with foam
cannons which are mounted on their cannons.
Aerial firefighting began around 1920 with first
attempts at dropping water tanks onto a fire. Most of these
attempts were unsuccessful during that duration, but they
provided a platform for initiatives in the future. In 1935, the
Aerial Fire Control experimental Project was created. At this
point, aircraft became important for fire detection. Inspired by
agricultural spraying techniques, helitanker soon became the
next progression in aerial firefighting. With firefighting
helicopters came heli-rapellers firefighters who rappelled
down to the fire as a ground force. In the 1960s, surplus military
aircraft became another effective way to suppress wildfires.
Aircrafts like the B-25, Douglas B-26 and Lockheed PSV were
able to carry 1000 gallons of water. Similarly some helicopters
use fire-watch cameras for fire surveillance. Its infrared thermal
147
imager can detect heat of a wild fire even through thick smoke.
The aim of this project is to carry out feasibility study of
installation of firefighting equipment on MI-171 helicopter and
determine its practicability. A study will be carried out that
whether MI-171 can be used for firefighting purposes. By virtue
of its qualities MI-171 is successfully operated in different
conditions whether it is heat or cold, rain or snow, mountings
or desert, sea or land. It has wide range of mission
accomplishments like transportation of up to 4000 kg cargo
inside cargo compartment, transportation of up to 4000 kg
cargo on external sling, SAR missions, fire-fighting, patrolling,
construction works, offshore oil rig flights, ambulance mission
etc. a design of the firefighting equipment, that is to be installed
will be given first, and then the whole design will be analyzed.
Proceedings
and bending loads formers were integrated on internal skin of

the water tank. Figure 1: Helitanker
METHODOLOGY
Figure 1: Helitanker
Firefighting helicopters are classified in two

categories based on firefighting mechanism and apparatus
Helitanker
Bambi bucket
Now to install this equipment on the floor of cargo

compartment, there are mounts available on it for the auxiliary
fuel tanks. Auxiliary fuel tanks are extra fuel tanks which are
carried by the helicopters to increase the range and endurance.
Once the fuel from the main fuel tanks is consumed completely,
then these are connected to the engine. Generally the volume
capacity of a single tank is almost 8003 . A cargo helicopter
can carry 4 auxiliary fuel tanks.
But the distances among them are according to the
dimension of the stands of auxiliary fuel tanks. A model
simulating the floor of the aircraft was generated in CATIA
software. The hard points are prominent in the following
diagram. Figure 2: Floor of Cargo Compartment
Both have their own advantages and disadvantages which are

as follows
Firefighting with helitanker aircraft speed limits are
not that much affect while Bambi bucket has the
limitation of maximum speed i.e.
o 80 120 kph
(with empty bambi bucket)
o 60 180 kph
(with filled bambi bucket)
o 60 80 kph
(while dumping water)
If we exceed these limits the bucket starts swinging
forth and backward, and even it can touch the tail rotor
of the helicopter
During operation aircraft CG is not disturbed in case
of helitanker while the CG shifts in case of bambi
bucket
Extra care is required during take-off and landing
In Bambi bucket, the fire retardant can only be
concentrated at the point exactly below the opening of
the bucket, no other direction is possible, as it falls
with action of gravity
Based on the above discussion we selected to operate from
a water tank (helitanker) installed the cargo section of our
Helicopter. Now for this first of all the volume and weight
capacity of cargo compartment were evaluated from manuals of
the aircraft. Volume capacity of helicopter 12.5 3 and weight
capacity is 4000 kg. First of all a water tank was designed in
CATIA software. Its volume capacity was 32003 . To avoid
column buckling and resist the circumferential stresses, bulk
heads and stringers were integrated with internal skin of the
water tank. These bulk heads are having hole so that water can
easily pass in between the section of the tank. These bulk head
structures also avoid sloshing effect. Similarly to resist axial
148
Figure 2: Floor of Cargo Compartment
So a new stand was designed according to the dimensions of

helitanker. The stand was installed on a layer of CFRP sheet
which has the mounts in accordance with dimensions of cargo
section. Figure 3: Stand
Proceedings
helicopters also. In addition to its advantage of low weight, it is

axially balanced with snorkel, and thus easily supported and inline with axis of the snorkel. This electric pump can deliver
1000 gpm through a 6 diameter snorkel. The snorkel length
varies from 12 to 15 feet. It operates on 7.5 horsepower. It will
draw this power from onboard generator. In this invention, an
electrical power cable extends from the electrical source of
helicopter along the outside of the snorkel and goes into the
protective shrouding on the pump unit. Once the pump is
lowered into the water reservoirs such as lakes, swimming
pools etc. the shrouding protects the pump unit from any
hazardous material such as rocks or wall of pool during filling.
Figure 3: Stand
Water will be released from emergency hatch of cargo
aircraft which will go down with action of gravity. For this
purpose an opening is provided in the CFRP sheet. This sheet
will fix on the floor with the help of screws just as the stands
for auxiliary fuel tanks are fixed. The dimensions and the
distance between the threaded holes is in accordance with hard
points available on the floor. Similarly the stand to hold the
water tank is fixed on the same sheet. This whole package when
combined together completes the structural design of the
firefighting equipment. This whole structure is fixed on the hard
points already available for auxiliary fuel tanks. Figure 4:
Complete Design. As our main focus is on the weight constraint
during the design phase so a high strength to weight ratio
composite CFRP is selected for manufacturing and further
analysis. Material properties will be computed later on.
Figure 5: Snorkel Pump

Now to spray the fire retardant on the burning fire,
doors are extended from the helitanker through emergency
hatch of the floor. The maximum flow rate from this opening
will be 1.19 m3/s. Once the doors are opened, the fire retardant
will fall directly from the helitanker with action of gravity.
These doors are controlled by microprocessor which actually
adjusts for airspeed and allow the doors to open for the flow of
water. The flow rate matches the particular coverage selected
by the pilot himself. Figure 6: Helitanker
Figure 4: Complete Design

For pumping of water, an Electric In-Line axial flow
pump will be used in whish an axial flow pump is mounted on
the free end of a snorkel which is fixed on the onboard storage
tank on the other end. Figure 5: Snorkel Pump. We could also
use the hydraulic centrifugal pump but then we would have
faced some weight penalty. Similarly the low power
requirement of electrical motor makes it suitable for small
149
Figure 6: Helitanker
Proceedings
= 17.61
MATERIAL PROPERTIES
For analysis, material properties of CFRP were
calculated. To have symmetric properties in all the three
direction, Quasiisotropic (0, 90, +45, and 45)Figure 7: Stacking
Sequence sequence was selected as stacking sequence for our
material
Similarly becomes
= 3.60
To find out the elastic modulus we will find out the A
matrix for above sequence
For this we need Q matrix for each ply
11
= [21
61
12 16
22 26 ]
62 66
11 =
1
11 = 134.5
22 =
1
22 = 17.7
12 = 4.619
66 = = 3.69
Figure 7: Stacking Sequence
To find out the properties of CFRP, ANALYSIS
AND DPERFORMAANCE OF FIBRE CCOMPOSISTEES,
written by BHAGWWAN D. AGAARWWAL was concerned,
in CFRP carbon is fiber while epoxy was as matrix, following
calculations were carried out
The expression for elastic modulus is as
The Q matrix becomes for 0 deg ply becomes

134.5
= [ 4.6
0
4.6
17.75
0
0
0]
3.6
For lamina at any angle the Q matrix becomes

11
= [21
61
= +
12
22
62
16
26 ]
66
We have
= 220
= 3.54
= 0.5
= 0.5
By putting these values
= 220 0.5 + 3.54 0.5
= 112
Now to find out elastic modulus in other two direction
1 +
=
1
Putting values in above eqn we get
Above values can be calculated from eqns as follows

11 = 11 4 +22 4 + 2(12
+ 266 ) 2 2
4
22 = 11 +22 4 + 2(12
+ 266 ) 2 2
4
12 = 12 ( + 4 ) + (11 + 12
466 ) 2 2
4
66 = 66 ( + 4 ) + (11 + 22 212
266 ) 2 2
66 = (11 12 266 ) 3 sin + (22 12
266 ) 3 cos
66 = (11 12 266 )3 cos + (22 12
266 ) 3 sin
Q matrix for 45 deg lamina becomes
44.14
= [36.76
29.09
= .95
150
36.76
43.96
29.09
29.09
29.09]
3.6
Proceedings
= 19.68
Q matrix for -45 deg lamina becomes
44.14
= 36.76
29.09
36.76
43.96
29.09
= 0.3449
29.09
29.09
35.76
= 0.3439
Q matrix for 90 deg lamina becomes
17.75
= [ 4.6
0
4.6
134.5
0
0
0]
3.6
Now material properties can be found as
The above calculated values were verified by using

NASTRAN software. Figure 8: Nastran Pattern Results These
values will be embedded in ANSYS Workbench as engineering
data in orthotropic properties. Further computation will be
carried out in the next steps.
2
11 22 12
22
2
11 22 12
=
11
12
=
22
12
=
11
66
=
So from above expressions we found out the matrix as
follows
Total thickness is 10 mm & number of lamiae are 20, so the
total thickness of each angled ply becomes 2.5 mm
[] = 2.5 [[ ]0 + [ ]90 + []45 + [ ]45 ]
Putting all the related values in the above equation we get the
A matrix as
601.25
= [ 206.8
0
206.8
600.4
0
Figure 8: Nastran Pattern Results
0
0 ]
196.8
Putting the values from above in the expression of Ex, Ey, Vxy,
Vxy & Gxy, we get
= 53
= 53
COMPUTATIONAL ANALYSIS
To check out for the structural integrity of the design, this

CATIA file was converted in IGES file. This igs file was
imported in ANSYS workbench. The static structural analysis
was carried out using different conditions. First of all the water
tank, structural analysis of each part is carried out separately.
The design of water alone was analyzed. The water of pressure
151
Proceedings
was applied on internal skin of the water tank. The pressure

applied was 2565 pa. Now to carry out these simulations fine
meshing was created. After that, the structure was put on
simulations. The following result was obtained.Figure 9:
Simple Analysis of Helitanker
Figure 11: Static Analysis of Complete Structure
Figure 9: Simple Analysis of Helitanker

The values shown along the colored graph are VonMises or equivalent stresses. According to these results the
maximum stress lies on the internal corner of the opening of the
tank. This maximum stress is in quite safe limits. Similarly the
frame was analyzed separately and the force was applied which
would be applied on it when the water tank is filled. Figure 10:
Static Structural Analysis of Stand
The results obtained still shows that the point of

maximum stress is located at the corner of the door. The
maximum Von- Misses stress still under the safe
considerations. Now to carry out the analysis while the aircraft
is in the move climb or any bank, values of gs was embedded
in the initial conditions. For this, first acceleration was
converted into inertial loads and this inertial load was applied
as the nodal force on the bodies with pink color in the following
diagramFigure 12: initial Conditions for Dynamic Analysis
Figure 12: initial Conditions for Dynamic Analysis
Figure 10: Static Structural Analysis of Stand

From the above results, still it is evident that maximum
point of maximum equivalent stress, the material will not fail.
After these simulations, the whole design, completely
assembled, was analyzed in the static position (Helicopter is at
rest with tank filled with water). The forces applied were
changed into pressure form because these forces were
distributed across the surfaces. Figure 11: Static Analysis of
Complete Structure
152
Proceedings
After running the simulations, following results were

given by ANSYS workbench. Values in the green boxes shows
Von Misses stress at that location.Figure 13: Results from 4g
Loads
Figure 15: Deformation at 5g Loads
Figure 13: Results from 4g Loads
Frrom the results of total deformation the maximum

deformation, the maximum deformation is 0.11176 m which is
too high and strucure will eventually break. So with this design,
we have to keep ouselve within 4g limit. We must try to stay
maximum at 3g. this holds good for all manuvers, translations
and rolls.
The maximum equivalent stress shows an abrupt rise

due to the acceleration. Similarly the total maximum
deformation was too high. Similar analysis was carried out on
inertial load with respect to 5g acceleration. The following
results were obtained.
When simulations for 5g inertial loads were applied on
the whole structure, the maximum equivalent stress increased
by a big amount. This value is still under the elastic limit range,
so the design will not fail. Figure 14: Results of 5g Loads
CONCLUSION:
With this we can conclude that any cargo helicopter
with weight lifting capacity of 4000 kg or more can be modified
so that it can be used in Aerial firefighting. For that a helitanker
is installed inside the cargo section of the aircraft, which will
by a snorkel from the water reservoirs easily. If we use light
composites like CFRP, it will have enough strength to resist
flight variations up to 3g. The helitanker will have the capacity
to carry 3200 liters of water. Compared with firefighting
helicopters, it will be an efficient helicopter.
[1]
[2]
[3]
Figure 14: Results of 5g Loads
[4]
153
REFERENCES
F. Qian, V. Strusevich, I. Gribkovskaia, and .
Halskau, "Minimization of passenger takeoff and
landing risk in offshore helicopter transportation:
Models, approaches and analysis," Omega, vol. 51, pp.
93-106, 2015.
D. G. Nichols Sr, "Electric in-line snorkel pump for
helicopter tanker and method of operation," ed:
Google Patents, 2001.
E. G. Keating, A. R. Morral, C. C. Price, D. Woods,
D. M. Norton, C. Panis, et al., Air Attack Against
Wildfires: Rand Corporation, 2012.
B. D. Agarwal, L. J. Broutman, and K.
Chandrashekhara, Analysis and performance of fiber
composites: John Wiley & Sons, 2006.
Proceedings
Design and Optimization of S-Band Wilkinson Power

Divider for Transceiver Applications
Muhammad Saad Sohail1, Talha Khan2
1, 2,
Space and Upper Atmosphere Research Commission (SUPARCO), Karachi, Pakistan

1
2
saad435@hotmail.com, aquarius_132@hotmail.com
AbstractThis paper presents a high performance and

compact size single band Wilkinson Power divider for wireless
applications. Basic three-port equations are used for designing
the Wilkinson at the operating frequency range, further tuning is
done by converting the lengths and impedances of each section of
conventional Wilkinson power divider into single band T-shaped
sections. ADS 2011 is used for the schematic design and EM
simulations are performed using Momentum. The reflection
coefficients are analyzed at the frequency range of 22.25GHz.The computed and the measured S-parameters are in a
very good agreement.
However, substrate cost is the leading aspect for micro strip
passive circuits, there is a requirement to make use of cheaper
materials such as FR4 without compromising the performance.
In this paper, FR4 has been used as a substrate material for the
development of Wilkinson power dividers operating in portions
of the S-band (2.5-2 GHz).
KeywordsMicro strip, T section, Wilkinson Power Divider,
Optimize Transceiver.
Design for center frequency of 2-2.25 GHz and Z0 = 50

requires the isolation resistor to be 2Z0= 100 and the
impedance of the quarter-lambda transmission line split
section to be 2 Z0 = 70.70 .
Fig.1.Impedance description of Wilkinson power divider
I. INTRODUCTION
Wilkinson power divider is a commonly used device in
microwave applications. Quarter wave transformers are used in
Wilkinson power dividers which are easy to develop on printed
circuit board using inexpensive method whereas still providing
high level of performance. As PCB transmission line is
employed, lumped and micro-strip line can be used for
accuracy in results. The designed power dividers are simple in
approach yet provide excellent performance in terms of
insertion loss, port matching and isolation.
WPD divides input power into two output ports. In ideal
conditions, this split is lossless. The divider requires all ports to
be matched. Since, three port are never reciprocal and matched
at the same time, a resistor is introduced which absorbs energy
due to any mismatch. This resistor also isolates output ports
which makes this device to also be used as a power combiner.
S-parameters are measured by applying signal to each port.

The reflection power waves are measured at remaining ports.
Calculation of S-parameters is the ratio between incident and
reflected power waves.
S-parameter for an ideal Wilkinson Power Divider has the

following properties:
All ports are matched (S11, S22 and S33 equals 0).
Ports 2 and 3 are isolated (S23 and S32 equals 0)
The power present at port 1 is equally divided between
ports 2 and 3 (S12, S13, S21and S31 equals -3 dB with a
90 phase shift)
The circuit is reciprocal (Sij equals Sji)
II. BRIEF INTRODUCTION TO WILKINSON POWER DIVIDER
III. DESIGN
Loss less power dividers can be a formed a WPD if the

output ports are matched. Reflected power is dissipated. Input
power can be split into two or more in-phase signals with the
same amplitude.
Micro-strip based equal split WPD is feasible to make without
any lumped components. The design considered in this paper is
micro strip based, as shown below in Fig. 1. The equivalent
transmission line circuit is shown in Fig. 1.
Initial design problem was to analyze a non-optimum

Wilkinson Power Divider and then to optimize it using Agilent
ADS and produce the usual values for line impedance (Z) and
isolation resistance (R). For this task, the single stage
Wilkinson Power Divider was implemented as shown in figure
(3) using the ideal transmission line model.
To determine the initial values of line impedance (Z) and
isolation resistance (R), two variables were defined as: r1 = 50
154
opt {10 to 500 }, z1 = 50 opt {20 to 150} The

nominal values of r1 and z1 were specified to be 50 with r1
ranging between 10 to 500 and z1 ranging between 20
to 150 . The nominal value was defined to be taken as the
seed value whereas the range was defined as an optimization
window. The simulation settings as mentioned below were set
to obtain the S parameters,
Start = 1.0 GHz; Step = 250 MHz; Stop = 4.0 GHz
The Optimization goals were set as: dB (S11) < 35; dB(S13)
= 3.01 and dB(S12) < 50.
LineCalc tool of ADS 2011 was used to convert the ideal
micro strip lines to the practical values as shown in Figure 2.
Proceedings
The results were slightly deviated due to EM effects.

Therefore, co-simulation was performed to optimize the
results.
Fig.4. Layout design of final Wilkinson Power divider
IV. SIMULATION AND COMPARISON RESULTS

The design was the simulated and optimized on FR4 and
AD450 substrates for comparison reasons having same
thickness and relative permittivity but different loss tangents.
The various properties of the substrates used are shown in
following tables.
Substrate parameters for Fr4
Fig.2. LineCalc tool of ADS for converting ideal transmission lines to micro
strip line
Quarter wave transmission line (g/4) is designed the required

frequency. The length g/4 could be folded using the curved
structure to reduce the size of design as shown in Fig. 3.
Er
T
Rough
4.5
17um
0mm
H
Cond
TanD
0.508mm
5.88e7
0.016
H
Cond
TanD
0.508mm
5.88e7
0.0045
Substrate parameters for AD450
Er
T
Rough
4.5
17um
0mm
The final measurement results of S-parameters as obtained

from Vector Network Analyzer (VNA) for both the substrates
are compared in Figures 5, 6, 7 and 8:
Fig.3. Schematic design of Wilkinson Power divider
Layout was generated as shown in Figure 4 using Fr4

substrate and EM simulations were performed in Momentum.
155
Proceedings
-22
-24
dB(S(2,2))
dB(S(1,1))
-26
-28
-30
-32
-34
-36
-38
2.00
2.05
2.10
2.15
2.20
2.25
2.30
2.35
2.40
2.45
2.50
freq, GHz
Fig.5. Reflection parameters using AD450 substrate
m1
freq=2.250GHz
dB(S(2,1))=-3.272
m2
freq=2.250GHz
dB(S(3,3))=-36.784
m1
Fig.8.Reflection parameters using FR4 substrate
-5
As can be seen from the results of above plots, that

performance of Wilkinson power divider is almost same for
both the substrates i.e. low cost FR4 substrate and high cost
and low loss AD450 substrate. Therefore, it can be said that
the design fits well for any substrate and is very useful for the
cheaper substrates without compromising the performance.
dB(S(3,3))
dB(S(2,1))
-10
-15
-20
-25
-30
m2
-35
-40
2.00
2.05
2.10
2.15
2.20
2.25
2.30
freq, GHz
Fig 6.Reflection parameters using AD450 substrate
2.35
2.40
2.45
V. CONCLUSION
2.50
RF transceiver commonly have a Wilkinson power divider

(WPD) or combiner circuit, techniques for designing WPD are
diverse. The paper shows comparison of WPD design on two
substrates having different loss tangent. FR4 being cheap and
easily available is defined as equally comparable to AD450 in
simulations. Next step is the fabrication and testing of these
WPD. The final circuit occupies less space and has good
matching results. With the thickness of 0.5mm, FR4 has
comparable behavior to specialized substrates and can be
considered as a robust and cost effective alternative for
applications in S-band.
REFERENCES
[1] Cripps, S.C., RF Power Amplifiers for Wireless Communications 2nded,
Norwood,MA: Artech House, 2006
[2] David M. Pozar, Microwave Engineering, John Wiley & Sons Inc., 2005
[3]Guillermo Gonzalez,Microwave Transistor Amplifer Analysis and design,
Prentice Hall
[4] Maas, Stephen A.Nonlinear microwave and RF circuits, Artech House
microwave library, 1997.
Fig.7.Reflection parameters using FR4 substrate
156
Proceedings
OPTIMIZE MANUFACTURING OF
UNIDIRECTIONAL CARBON PREPREGS FOR
SPACE APPLICATIONS
Mateen Tariq Ahmad Bashir, Dr. Sohaib Akbar, Dr. Sajid Mirza
Pakistan Space and Upper Atmosphere Research Commission
Karachi, Pakistan
Mateen_90@hotmail.com
Abstract High performance composite prepregs are used

globally in aerospace applications for faster and easier
manufacturing.
Manufacturing
composite
parts
using
unidirectional carbon prepregs offers flexibility to optimize the
design and layers, resulting in superior properties. Using
prepregs as manufacturing medium, parts can be manufactured
with low void content. Unidirectional prepregs give control over
thickness hence high specific modulus and strength can be
achieved. Fatigue and acoustic performance of composite is
strictly related to type of reinforcement. Unidirectional
reinforcement plies offer good fatigue and acoustic performance
making them suitable candidates for space manufacturing.
Superior properties make them viable for manufacturing of
various parts of Launch vehicles such as payload fairing, motor
casings, fuel tanks and shims for flexible nozzle assembly. This
paper deals with the in-house manufacturing of unidirectional
carbon prepregs. The prepregs were manufactured using carbon
fiber tows and epoxy based prepreg resins. The fiber from tow
was mechanically placed parallel to each other resulting in
unidirectional carbon fiber prepregs. A solvent dip process using
resin bath and roller was used to manufacture prepregs. To stock
the batch of prepreg, cold storage system was utilized. The
process was developed using customized machinery and
equipment. Tests on prepregs showed satisfactory results in
terms of volume and weight fractions. Finally the prepregs were
cured and formed into unidirectional sheets for further tests.
Prepregs were cured in autoclave utilizing vacuum bagging
operation. The cured sheets were subsequently tested which
showed competent results hence asserting the material as suitable
candidate for use in space applications.
Index Terms Unidirectional Carbon, Prepregs, Polymer
matrix composites.
I. INTRODUCTION
Space applications demand materials with high specific
stiffness and strength so that weight of structure may be
minimized. This need of space industry has lead towards
development of many advance fiber reinforced composite
materials. With the development of light weight composite
structures, many structural challenges can be addressed.
Among various advance composite materials, unidirectional
fiber reinforced composites offer better control on properties
[1]. With the use of unidirectional laminates, desired strength
can be achieved by placement of laminates as be modeled
design. Moreover degree of anisotropic behavior may also be

altered using unidirectional laminates.
Prepreg technology allows fabrication of intricate part with
high quality. Prepregs may be used manually as in hand layup
process of with automated placement machines. For
manufacturing of parts for space applications consistent and
high quality is required which can be achieved by using
prepregs. The desired volume fraction of fiber for prepreg for
space application is about 55-60% [2]. Choice of matrix and
reinforcement depends on the application. Carbon fiber
impregnated with epoxy matrix is used in many applications as
this combination offers superior mechanical properties.
Carbon composites with epoxy matrix have been used
extensively in producing aerospace parts such as landing gear
doors, flaps, aileron and other structural parts. Unidirectional
carbon fiber- epoxy composites offer competitive fatigue
properties in low earth orbits [3]. With superior properties,
unidirectional carbon fiber composites may be used as face
sheets for satellite payload fairing where high strength and high
fatigue life is required. As the space environment offers
vacuum thermal cycling, superior thermos-mechanical
properties of unidirectional carbon fiber composites make them
viable candidate for manufacturing various satellite parts
This paper is distributed in two parts, manufacturing and
testing. The former explains the in-house manufacturing of
carbon fiber prepregs and the subsequent sheet forming,
whereas the later describes the testing results of manufactured
unidirectional carbon fiber sheets.
II. EXPERIMENTAL
A. Materials
Material used for this study were oriented towards high
specific strength. Carbon fiber tow (10K) was used as
reinforcement for prepreg forming. Carbon fiber was chosen
due to its high strength and low coefficient of thermal
expansion. Choice of resin system depends upon the
application hence for high strength application, epoxy based
resin is desired. For this study, Swancor 2552 prepreg resin
was chosen due to its low viscosity and good impregnation
properties.
157
B. Manufacturing of prepregs
Manufacturing of prepregs was performed using in-house
developed fiber impregnation machine. The scheme of
manufacturing involves impregnation of dry fiber and then
placement of fibers parallel to each other to form a continuous
sheet.
The manufacturing process of prepregs is somewhat related
to the process of filament winding. Dry fiber was firstly
impregnated with resin using a drum-roller mechanism same as
for filament winding process. The schematics of impregnation
of fiber is given in fig 1.
The open surface of prepreg was covered with non-stick

paper and prepreg was wrapped in roll form. Afterwards, the
roll was sent to cold storage to store until use. The specification
of prepregs and process parameters is given in table1.
TABLE I. PROCESS PARAMETERS AND SPECIFICATIONS
Fig. 1: Schematic for Impregnation of Carbon Fiber Tow
After impregnation of fiber with prepreg resin, the fibers is

passed through pressing rollers which align the filaments in
longitudinal direction and avoid any overlap of filaments
within fiber layer. This process is desired because it flattens the
fiber and aligns it so that perfectly straight and low thickness
sheet of prepreg may be achieved.
The last stage of prepreg setup consists of a large rotating
mandrel, 2400 mm in diameter containing a release agent
sprayed non-stick paper. The wet fiber was wound over the
mandrel in such a manner that neither the two parallel fibers
overlap each other nor there a gap in between them. This
perfect alignment of fibers side by side is achieved by
controlling the operating parameters with respect to band width
of fiber.
Subsequently after the desired width of wound has been
achieved, the fibers were cut in circumferential direction and
unwound so that flat sheet of prepreg is formed. Fig 2 shows a
prepreg sheet.
Proceedings
Fiber
Carbon fiber 12K
Matrix
Epoxy-Swancor 2552
Width
355 mm
Length
6000 mm
No. of Spools
01
Roller Speed
36.9 mm/min
Dry Fiber Weight
32 g
Releasing agent
Mold-Viz Spray
C. Curing of prepreg sheets

Prepregs were drawn from cold storage for manufacturing
of composite structure. Prepregs were unwrapped and
protective covering was removed from one side. Placement of
prepregs on mold is a intricate job. As the Fibers are not linked
with each other mechanically, it becomes difficult to place a
laminate of unidirectional fiber on surface. Back non-stick
paper layer acts as a supporting medium to bind the fibers
together. Fiber were placed on mold manually with hand-layup
and afterwards, paper layer was removed. The placement
direction of fibers depend upon the design of structure. The
sample for mechanical testing was prepared with unidirectional
layer giving its maximum properties in one direction.
Curing of unidirectional sheet was carried out in autoclave
using air pressure and temperature up to 150 oC. The whole
process was carried out in 2-3 hours resulting in rigid
unidirectional carbon fiber reinforced epoxy sheet. Fig 3 shows
a cured sample with unidirectional laminates.
Fig. 3: Cured unidirectional carbon fiber sheet
Fig. 2: Prepreg sheet
158
D. Testing of Composite samples

Two type of tests were carried out in this study. Initially,
prepregs were tested for calculating resin content. Four
samples of prepregs before curing were tested to calculate
average value.
Mechanical testing was performed on cured sheet to obtain
strength parameters. Tensile test according to standard ASTMD 3039 [4] was performed on a universal testing machine.
III. RESULTS AND DISCUSSION
A. Resin Contents
Resin contents were calculated in terms of total weight.
Tests results indicates average value of 40% resin content. Also
the presence of solvent was calculated as approximately 3%
within the matrix system.
REFERENCES
[1] Purslow D., The shear properties of unidirectional carbon fiber
reinforced plastics and their experimental determination,
Aeronautical research council current papers, 1977 Retrieved
from: naca.central.cranfield.ac.uk/reports/arc/cp/1381.pdf
[2] Larberg Ylva, Deformability of Unidirectional Prepreg
Materials, Licentiate Thesis Stockholm, Sweden. 2009
Retrieved from: www.diva-portal.org/smash/get/diva2:224958/
FULLTEXT01.pdf
[3] A. Anvari, Fatigue Life Prediction Of Unidirectional Carbon
Fiber/Epoxy Composite In Earth Orbit, Int. J. of Appl. Math
and Mech. 10 (5): 58-85, 2014 Retrieved from:
ijamm.bc.cityu.edu.hk/ijamm/outbox/Y2014V10N5P58C25132
388.pdf
[4] ASTM D 3039, Standard Test Method for Tensile Properties of
Polymer Matrix Composite Materials,
B. Mechanical testing
Tensile test was performed on cured unidirectional carbon
samples. The average results are summarized in table 2.
TABLE II. RESULTS OF MECHANICAL TESTING
Breaking load
23 KN
Ultimate Tensile Strength
1440 MPa
Modulus of elasticity
97 GPa
Elongation
1.4 %
Proceedings
IV. CONCLUSION
Unidirectional carbon fiber prepregs were manufactured
using in-house designed prepreg machine. The prepreg
machine was designed on the concept of filament winding. Dry
fiber from spool was impregnated in a resin bath and then
wrapped around a large mandrel. Semi cured prepreg sheets
were removed from mandrel and stored in low temperature
environment for preservation. The sheets were then withdrawn
and cured in unidirectional laminates to form rigid sheet for
subsequent testing. Test results of prepreg sheet and cured
composite part shows satisfactory results. Prepregs contain
40% resin which is a nominal value. Moreover, cured sheets
offered excellent mechanical properties. Hence composite
structure with light weight and high strength was formed using
unidirectional carbon fiber prepregs.
159
Proceedings
Electrical Power Conditioning Unit Design for Space

Qualified C-Band Receiver in GEO Satellite
Applications
Muhammad Azeem Aslam
Beihang University, (BUAA)
Beijing, China
Ms.azeem@gmail.com
Taimoor Zahid
Institute of Space Technology, (I.S.T)
Islamabad, Pakistan
Taimoor187@yahoo.com
Abstract The competition in commercial satellite market

requires the Electrical Power Subsystem (E.P.S) components to
be more and more adaptive to the payload specific
characteristics. The Electrical Power Conditioning Unit (EPC) is
an important part of GEO Satellites, especially in payload
receivers. Because of input/output voltage, reliability, antiradiation, in-rush current protection, TM/TC, EMC, bus voltage
ripples and many more requirements which makes its design
complex and lengthy for an RF point of view. Apart from these
design driven factors the design should also be cost effective and
mechanically stable. The C-band microwave frequencies have
excellent performance in most physical aspects like rain fade,
atmospheric attenuation at low elevation angles and cheaper
bandwidth. This aim of this paper is to describe the functions and
design of Electronic Power Conditioning (EPC) module for a CBand Receiver for a given set of space qualified requirements. To
fulfil the above said requirements a new generation of EPC
design is presented by implementing half bridge topology and
more efficient rectification and regulation techniques. The output
voltage ripples have been minimized to 25mV and input filter is
designed by using Differential Mode (DM) and Common Mode
(CM) filtering techniques to overcome EMC/EMI interference.
Under voltage and over current protections are also implemented
to protect satellite in case of short circuits or lower main bus
voltage. De-rating analysis is performed by considering the
ECSSQST3011C space standard.
It also describes the
compliance with performance requirements for RF, electrical
and mechanical interface during the EPC module design for CBand Receiver. The design presented in this paper complies with
performance requirements for RF, electrical and mechanical
interface. Considering all the above mentioned factors the design,
simulation results and analyses are presented in paper.
interface during the EPC module design for C-Band

Receiver. The main functions of EPC unit are:
Conversion of DC bus Voltage to a different high
stability output Voltage.
Controlling the sequence between positive voltage
and negative voltage.
Low output voltage-ripples.
Perform TC ON/OFF, Telemetry etc.
Perform over-current protection, under-voltage
protection.
Fulfil EMC requirements.
Fulfil high reliability requirement.
For the power bus voltage there is an operational and a nonoperational requirement. The operational voltage ranges are
given below:
Power bus voltage range of the normal work mode is
90110 (Nominal value is 100V).
The safety power bus voltage range8590V and
110115V
In the normal voltage range of the power bus, the equipment
meets all performance requirements. In the safety voltage
range of power bus (8590V, 110115V), basic equipment
functions keep working.
The non-operational power bus DC voltage range is 0V-115V.
In the non-operational DC voltage range, the equipment
should not be damaged. Variation slope is less than 10V/msec.
The output current and voltage requirements are given in the
table below:
V1
Index Terms EPC, GEO, EPS, and EMC/EMI.

Voltage or Voltage Range
I. INTRODUCTION
The Electronic/Electrical Power Conditioner (EPC) is
actually a multiple output DC-DC converter which provides
power to a TWTA or an SSPA from the satellite power bus.
The power provided to both amplifiers is normally from
more than one source and is well regulated. The power
requirement depends on the efficiency of the amplifier and
transmitter power requirement. This aim of this paper is to
describe the functions and design of EPC module for CBand Receiver. It also describes the compliance with
performance requirements for RF, electrical and mechanical
(V)
V3
Remarks
+12after
-5
+6
CW7812
Voltage Tolerance V (V)
0.2
0.2
0.2
RippleP-P(mV)
50
50
50
5-15
580-810
Current or
SS in
current
vacuum
rangemA
SS in
ambient
160
V2
165230
In this
range,
first
5-15
750-1050
165-
value is
230
nominal,
second
Conducted Susceptibility
5-15
165-
1000-1400
230
min)
Appropriate suppress switching transient
by
Conducted Emission
adding
regulator block
CW79L05
CW7812
Whether the negative

voltage is first out when
switch ON and negative
Yes
voltage last when switch

OFF
Table.1 Output Current and Voltage Requirements
The equipment will be commanded ON/OFF through pulse

signal. Command signal is a positive voltage pulse. It is
necessary to take anti-interference measures for the input to
perform remote control command. The TC circuits will not
operate when supplied with a pulse of 100s duration, up to
5V amplitude and a repetition rate of one per 13 msec. The
TC circuit cannot operate correctly when subjected to the
nominal pulse with one break period of 1sec which occur
at any time during the TC pulse period. The signal
waveform and timing are described in the figure below.
Figure.1 Telecomand signal waveform profile
The values of tr and tf be equal to or lower than 50s. The

telecommand requirements of the C-band receiver are given in
the table below:
Parameters
Pulse Voltage-True
State
Pulse Voltage-False
State
Requirements
10V ~ 12V
Table. 2 Telecommand Requirements
The Digital Telemetry (BL) features are:

1) Output Voltage
High state corresponding to logic "1" level: 10.0V~12.0V
Low state corresponds to logic "0" level0V~0.2V
Exceptions conditions do not exceed 0~+15V
2) Output impedance5k
3) Output short circuit protection
The short circuit of output and ground do not cause any
permanent damage and also do not affect the performance of
other parts of equipment.
With the interconnecting harness(es) disconnected, the
equipment should have a minimum DC resistance of 1M
between the power input leads and the equipment case. It
should also have secondary power supply circuits connected to
the equipment case. These circuits should be isolated from the
primary power circuits by a minimum DC resistance of 1M.
The equipment should have the command return isolated from
the power return and also from secondary ground. The
equipment must continue to operate normally without
degradation with either a short or open circuit on one or all of
telemetry output lines and must not be damaged by either
short or open circuit on the command or power lines. To avoid
failure, a fuse must be used to protect the main bus. The over
current protection circuit (fuse) must start/be used when bus
current goes up to 2A. Current consumption of the equipment
should be limited to twice the maximum steady state current
within 5ms. Receiver must incorporate under-voltage
protection circuitry that will automatically turn off the receiver
when the primary power input voltage falls below 85V. The
response time does not exceed 20ms.When bus voltage
reaches the nominal value, the equipment can be turned on by
telecommand only. The design of the equipment is done while
taking into account the consideration of the ground storage for
two years.
Failure Rate < 400 Fits @500C for the lifetime being 15 years.
The temperature should be as follows:
Operating temperature range-50C500C
Acceptance temperature range-100C550C
Start-up Temperature: -150C+650C.
Equipment boot inrush current requirements for design and
test (without evaluation) are shown in the figure below. I dc is a
maximum constant current under normal conditions.
0 ~ 0.2V
Pulse Duration
1002ms
Maximum current
< 1.2mA
Input Impedance
10K
due to inductive load which may cause the

driving source capabilities to be exceeded
40%
margin.
Requiring voltage
12V/100 s pulse width/10Hz
value is
Warm
up (10
Proceedings
161
Proceedings
requirements mentioned above have been fulfilled by this unit.

The result for different output voltage levels and output ripple
(50mV) has been simulated and can be seen from figures
below. These simulation results are before linear regulator
(LR), because it is difficult to model LR in P-Spice. LR is
used on 12V and -5V secondary lines. After LR the secondary
voltages will be at nominal values (12V,-5V and 6V). The
output ripples can be seen in figures below.
Figure.2 In-rush current Limits.
II.
MODULE DESIGN DESCRIPTION
EPC unit is subdivided into four major parts which are as
under:
Main DC-DC converter unit
Auxiliary DC-DC converter
TC Circuit
TM Circuit
Figure 4 Secondary 6V Ripples Level (20mV max.)
The block diagram of the EPC module is shown in figure

below:
Figure 5 Secondary 15V Ripples (35mV max.)
Figure.3 Block Diagram of EPC.
A. Main DC-DC Unit

The main function of this unit is to produce the output
voltages (-5V, 12V +6V) for C-band RF and the local
oscillator chain at different output power levels and 12V for
under voltage / over current protection (voltage used in
different modules on EPC PCB board). It is implemented
using half bridge topology using rectifier, output filter and
voltage regulator modules. The output current and voltage
B. Auxiliary DC-DC Converter

The main function of the auxiliary DC-DC converter is to
ensure the output voltage to be -5V. It has two outputs one is
15V and the other is -15V, -15V is for the output line of -5V
to make sure that -5V appears first and 15V is for the main
DC-DC unit to start it. The operating time is not more than
100ms because the TC ON pulse time is 100ms.The output
results are shown as under.
162
Proceedings
Figure.8 TC OFF Circuit Output
Figure 6. Auxiliary Converter Output (15V and -15V)
C. TC ON/OFF Circuit
This unit is used for switching on and off the EPC and all
output voltages only when Telecommand Unit (TCU) sends
command to EPC. This circuit generates 15V signal to
auxiliary converter PWM IC to start EPC and generates 5V
pulse for main DC-DC converter PWM to shut down EPC.
According to the requirement it can only be operated when
telecommand of 100ms is at the input of this unit. The TC
circuit does not operate when supplied with a pulse of 100s
duration, up to 5V amplitude and a repetition rate of one per
13 msec. The TC circuit operates correctly when subjected to
the nominal pulse with one break period of 1sec which
occurs at any time during the TC pulse period. The simulation
results can be seen from figures below.
TC OFF signal in figure above is at the input of main PWM

IC to turn it off. But when there is no TC OFF signal the
voltage across main PWM IC increases slowly. In real-time
circuit, the voltage across PWM cannot rise till TC ON signal
comes from the ground station. Hence The EPC shall be at off
state. When 12V TC ON signal appears at the input of
Auxiliary DC-DC converter it generates signal to main PWM
to start.
D. TM Circuit
This circuit provides a constant telemetry from 10 to 12V
when receiver is working and 0V when receiver is off. The
TM signal will be received for EPC when EPC is working.
When EPC is not working there will be no TM signal which
can be seen from the figure below.
Figure 9 Output Telemetry Signal
Figure.7 TC ON Circuit Output
III.
CONCLUSION
In this paper a design of the Electrical Power Conditioning

Unit (EPC) which is an important part of GEO Satellites,
especially in communication satellite payload receivers using
half bridge topology and more efficient rectification and
regulation techniques is presented. The aim of this paper is to
163
describe the design and analyze results of EPC module for

new generation communication satellite payload receivers for
a given set of specific space qualified requirements. The
output voltage ripples have been minimized to 25mV is
presented. The input/output voltages, in-rush current
protection, TM and TC characteristics, bus voltage ripples and
other requirements which make its design complex and
significant for RF point of view. Both the output current and
voltage requirements are presented and met successfully. The
design presented is also compliant with the TC requirements.
The under voltage and over current protections are also
implemented to protect satellite in case of short circuits or
lower main bus voltage. All the four major modules of EPC
(main DC-DC converter unit, auxiliary DC-DC converter, TC
circuit and TM circuit) are discussed and analyzed along with
their simulated results.
REFERENCES
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
Ioana-Monica, Pop-Calimanu, Prutianu Florin, and Popescu Viorel.

"Design and simulation of DC/DC boost converter used for distributed
sensing system based on a multidrop sensor network with RS485
interface", 2012 10th International Symposium on Electronics and
Telecommunications, 2012.
Pequet, E.; Delporte, P.; Fayt, P.; Gak, M.; Canon, T., "ESA qualified
EPC for telecommunication satellites TWTA," Vacuum Electronics
Conference, 2000. Abstracts. International , vol., no., pp.2 pp.,, 2-4 May
2000.
R. E. Sorace, V. S. Reinhardt, and S. A. Vaughn, High-speed digital-toRF converter, U.S. Patent 5 668 842, Sept. 16, 1997.
Castiaux, J.P.; Bury, P.; Liegeois, B., "Power conditioning units for high
power geostationary satellites," Power Electronics Specialists
Conference, 1997. PESC '97 Record., 28th Annual IEEE, vol.1, no.,
pp.722,733 vol.1, 22-27 Jun 1997.
http://www.space-airbusds.com
https://www.thalesgroup.com
http://www.sstl.co.uk
Garrigos, A.; Carrasco, J.A.; Blanes, J.M.; Sanchis-Kilders, E., "A
power conditioning unit for high power GEO satellites based on the
sequential switching shunt series regulator," Electrotechnical
Conference, 2006. MELECON 2006. IEEE Mediterranean , vol., no.,
pp.1186,1189, 16-19 May 2006.
164
Proceedings
Proceedings
DSP Based Electro-Hydraulic Actuator Control

with Irretraceable Feedback Error
Compensation
Najam-Ud-Din Ahmed, Syed Aleem Azhar, Khadim Hussain
AbstractThis paper describes the interfacing detail of

Electro-Hydraulic actuator with DSP and addresses the
issues which commonly occur during development. The
purpose of designing digital controller on DSP is its
easy configurability via software and requirement of
less analog part. Scenarios where a dc-dc converter
used as voltage source may induce erroneous feedback
in the feedback control system due to its dual output
mismatch. And despite a good PID controller real
position of manipulator may have errors which are
irretraceable even by the intelligent controller schemes.
A judicious and intelligent use of available DSP
peripherals like ADCs, DACs etc and minor adjustment
of circuit can effectively increase the real time accuracy
of whole system against irretraceable errors.
KeywordsDSP
based
PID,
PID
II.
PROBLEM STATEMENT
Two hydraulic pistons are used in a robotic arm

for high speed application prototype. Feedback is
based on linear potentiometer for which dual supply is
required. It is known that 1% error in supply voltage
will produce same error in the feedback which will be
not traceable in potentiometer based feedbacks and
still be added in the real position of actuator.
Output Voltage (Vout) at potentiometer viper is,
(1)
(2)
Controller,
Irretraceable error, Erroneous feedback compensation
I.
INTRODUCTION
It is common practice to use DSP or any digital

controller to control analog manipulators and
actuators in mechatronic applications [3]. However,
noise and supply voltage accuracy dependability of
analog systems is always associated with them and
affect the accuracy of the overall system. There are
many advanced as well as conventional control
schemes to increase the performance of the actuator
[2] but these schemes might be less affective if the
feedback is erroneous or we have high tolerances in
the supply voltages. This paper discusses such a
scenario where a dc-dc converter used as voltage
source may induce errors in the feedback system due
to its dual output mismatch. And despite a good
controller real position of manipulator may have
errors which are irretraceable by the intelligent
controller schemes. In addition, use of digital
controller specially a DSP helps in generation of real
time rate for interfacing. Although various techniques
are available for getting precise band-gap reference
voltage [1] but DSP based controller also adds noise
immunity which is always associated with analog
systems.
(3)
(4)
(5)
)
(6)
This error may be induced due to two factors

1.
Due to tolerance in supply
2.
Due to mismatch in dual supply
There is always a pinch of tolerance in supply and

it is always expensive to use high grade supply unit or
165
reference generation ICs. Thermal drifts associated

with the electronics circuits make this tolerance/error
always changing. Moreover, it is common observation
that plus side load is always greater than negative side
load of the supply or any DC-DC converters which
are dependent upon percentage load for its good
regulation. Hence, there is always supply mismatch
between its dual outputs. This condition becomes
more severe when there is a gain in the feedback loop.
Proceedings
Position Error (to be used in PID eq.) =

Setpoint - Non-compensated feedback position
This relation is valid where there is no mis-match
in 10volt references. And tolerance is zero or very
close to zero. This is not possible in real scenario.
Due to possible supply tolerances outside the
range; ADC will not be capable to measure the
voltage beyond +/-10V due to its saturation.
Therefore, +/-9.5V has been provided for reference
voltage instead of +/-10V to the potentiometer. This
will decrease the resolution of ADC by some amount
but provides the protection for the possible overflow
of reference voltage outside the ADC range.
Moreover, it will indirectly provide fault tolerance
against open/short connection of ADCs and feedback
potentiometers.
A current loop is added in the system to

compensate the noise affect associate with the analog
electro-hydraulic valve. An op-amp based current
loop will serve the purpose which was connected with
DSP DAC output.
Nonlinearity may also occur in feedback if the
feedback is just beyond the measurable range of the
ADC. For example, an ADC of 2.5 volt is being used
and feedback voltage of 2.6 volt will appear 2.5volt to
the controller algorithm and position corresponding to
0.1Volt increase will never be compensated. This
situation may occur due to saturation of ADC.
Now there
compensation.
are
two
errors
which
need
Full scale reference input error due to 9.5 volt

instead of 10 volt (instead of 20 it is 19 volt)
III.
HARDWARE DESCRIPTION AND

DSP ALGORITHM
Reference mismatch error due to tolerances in

+9.5 and -9.5 voltage references (creating offset)
DSP D-Module 21065L D-sign T is used for this

application along with S-module. In order to control
two actuators four ADC input and two DAC output
are used. Two ADCs are required for taking feedback
position of both actuators, two DAC output are
required to drive each actuator to the desired
calculated position. The ADC and DAC we used have
the 16-bit resolution and have different full-scale
voltage ranges of +/-2.5V, +/-5V and +/-10V. In order
to utilize full resolution of ADC the reference voltage
range provided to the actuators potentiometer
position sensor and the ADC full scale voltage range
must be matched. We select the ADC range and
correspondingly reference voltage range to +/-10V
because the external noise produces little impact on
this large value than using +/-5V or +/-2.5V range of
ADC. Two more ADCs are required for measuring
reference voltage supplied to the actuators
potentiometer for position sensing.
To compensate these errors we need to exactly

find the current reference voltages and then apply the
compensation as follows,
(
(
7)
((
) (
))
Now position error used in PID loop is Err

calculated as,
And desired output as,

Desired_Actuator_position=
A. Error compensations and implementation

The ADC available on S-module has 16-bit
resolution and has a bi-polar input of a range of +/10V. Actuator gives position feedback in the form of
varying resistance. To utilize the full resolution of
ADC, we have to supply +/-10V to the potentiometer.
Non-compensated feedback position =
input (1 or 2)
)
)
(
8)
Where, kp = proportional-gain, ki = integral-gain,

kd = differential- gain.
The tuning of digitally implemented PID has made
it much easier as compare to using of analog PID.
ADC
166
B. Saturation limit
A saturation limit is a safety measure must always
be applied in the algorithm to avoid overflow
conditions and generation of undesired & unknown
erroneous DAC output. For 16 bit DAC we have
saturation applied as,
Proceedings
IV.
DSP PROGRAM FLOWCHART
Main Program Loop

START
Initialize Dsp and Peripherals
If (Desired_Actuator_Position > 32767)

False
Desired_Actuator_Position = 32767
If (Time_Delay >
2msec)
ElseIf (Desired_Actuator_Position < -32767)
True
Desired_Actuator_Position = -32767
Prepare packet to transfer

serially
Else
False
Desired_Actuator_Position=
Desired_Actuator_Position
If (Transmit-Buffer = Empty)
True
Send Data-Packet
serially
Finally, the above calculated value is sent to the

corresponding DAC-Channel.
The scanning frequency of the above calculation is
done at the rate of 10 KHz. It means that ADC
sampling, the calculation for the desired output in PID
LOOP and updating on DAC output are all done at 10
KHz frequency. To ensure the constant sampling rate,
we do all the calculation at the 10 KHz TimerInterrupt.
If (60-sec Profile is not running &

Character is received at serial-port )
False
True
Receive Command
If (Command is for Sine-wave, Squarewave or for profile & Command is new)
C. Real time clock rate and RS422 Communication

In the 10 KHz loop, we also generate different
patterns of set-point on time-base. These set-points
contain sine-wave, square-wave, triangular-wave and
stair-case waveform. To evaluate the performance and
post-analysis of the above control we sent real-time
values of set-point and feedback along with the
absolute time in seconds in the 2msec interval to the
Rs-422 port of PC at the baud-rate of 115200.These
values are arranged in the packet form and have the
total length of 16-bytes.
True
Save absolute-time into offset-time and newcommand into previous-command
True
Reset Timer_Count to Zero.
END
Fig.1.
The above packet is transmitted by the DSP and

received by the application program built on the
Labview. As in the table shown above, the starting
two bytes used to synchronize the incoming packet
and the last byte contain the checksum of the data
bytes only excluding the synchronization byte. The
incoming packet considered to be valid if calculated
checksum and the received checksum contained in the
last byte of the packet matches. These values are
displayed in real-time in the application and also
logging the values in the text file for post-analysis.
167
Main program Loop
False
Proceedings
10khz Interrupt Loop
PID Function
START
Calculate absolute-time and
relative-time
START
Read Feedback position of Actuators and

Reference supply voltages using A/D
converter.
If (Command ==1)
Calculate offset generated by voltage-level

converter card by using the formula given below.
(+Vref)+(-Vref) / 2
True
Generate Set-Point(Profile) of both
actuators contain various waveform
of duration of 60-seconds.
If (Command ==2)
Remove offset-error generated from both

actuators feedback positions.
True
Generate Set-Point of both actuators
to value of 25mm.
Calculate Error for both actuators,

Error = Set-point Corrected Feedback
If (Command ==3)
True
to value of -25mm.
Apply Gain-Control on both actuators,

Desired-Position =
If (Command ==4)
(kp * Err) + (ki *
) + (kd *
True
to value of square-wave of 2Hz.
Apply Saturation limit to Actuator 1

desired-position to prevent overflow from
16-bit integer value.
(Because the DAC has 16-bit output)
If (Command ==5)
True
to value of sine-wave of 2Hz.
Apply Saturation limit to Actuator 2

desired-position to prevent overflow from
16-bit integer value.
(Because the DAC has 16-bit output)
If (Command ==6)
True
to value of sine-wave of 8Hz.
Output desired calculated position to

respective DAC channels.
If (Command ==7)
True
to zero millimiter(0 mm).
END
Call PID( ) function that perform on

both actuators position and also output
desired actuators position on DAC.
Fig.3.
END
Fig.2.
10kHz Interrupt Loop
168
PID Loop
V.
Proceedings
TEST PROFILE
A complete test profile is a good practice to test

the response of actuator. This profile depicts and
incorporates all or nearly all type of responses which
an actuator/ manipulator may experience during its
course of operation. A test profile for such purpose is
generated and preprogrammed in the DSP as shown in
fig 4.
Fig.6.
Fig.4.
Square Wave Response
Test Profile
This profile contains

1.
2Hz 25mm sine & square set point
2.
25mm Triangular set point
3.
Staircase Set point
4.
High frequency low amplitude set point
VI.
Fig.7. Step Input Response
RESULTS
Fig.8. Higher Frequency Low Amplitude Sine
Results clearly shows good dynamic response of

electro hydraulic system
VII.
Fig.5.
CONCLUSION
In this paper, we have utilized available ADCs and

high speed performance of DSP for compensation of
feedback errors. Use of extra available ADCs will
decrease the additional cost of high grade circuitry for
generation of highly accurate voltage references. A
simple resistive based feedback sensor will further
decrease the cost and interfacing effort in
analog/digital position feedback setups. Use of
already available ADCs will not only simplify the
Low Frequency high Amplitude Sine Wave
169
circuit but also the errors associated with use of

additional components in any circuit.
REFERENCES:
[1] LinHaiCui. Design of a High Precision Band-gap
Voltage Reference. Proceedings of the 2011
International Conference on Electronic & Mechanical
Engineering and Information Technology. IEEE, 2011 :
2187~2190.
[2] Sergey Edward Lyshevski and Trevor C Smith.
Tracking Control of Direct-Drive Servos. Proceeding of
the 2011 International Conference on Decision and
Control and European Control Conference(CDCECC).Orlando, FL, USA, IEEE, 2011 : 1602~1607.
[3]Hai-tao Wang, Ze-Zhang and Xiang-yu Liu. Design of
Control System for Brushless DC Motor based on
TMS320F28335.Proceeding of the 2011 Third
International Conference on Measuring Technology and
Mechatronics Automation. IEEE, 2011 : 954~958.
[4]D-module D-sign T manual.
[5]VisualDSP++ 4.0 - C/C++ Compiler and Library
Manual for SHARC Processors.
[6]ADSP-21065L SHARC DSP - User's Manual and
Technical Reference
170
Proceedings
Proceedings
Design for Test Approach using FPGA for BPSK

Modem
Syed Jahanzeb Hussain Pirzada1, Muhammad Faisal Arain1, Rahman Mehboob1
1
Space and Upper Atmosphere Research Commission (SUPARCO), Pakistan.
Abstract Implementation of modem has always been the key

interest for researchers working in the field of communication.
Nowadays, at low frequencies (in kilohertz range) the modulation
and demodulation is performed digitally. As, modem is utilized
for telemetry and telecommand communication between satellite
and ground station. This work provides the design of test bed for
modem. The testing is performed using ECSS based Standards.
This paper explains the design of test bed for testability of Binary
Phase Shift Keying modem implemented on Field Programmable
Gate Arrays. The testing of FPGA based modulator was
performed using signal analyzer. On the other hand the testing of
FPGA based demodulator was complicated and needs to be
design on seperate hardware; as it contains carrier recovery, bit
syncronization and bit recovery. Hence, digital signal processor
and FPGA based co-processing test bed was design and
developed for tesing of demodulator. The experimental results
show the architecture and testing results of the proposed
approach. The proposed methodology describes the problems
encountered during testing and proposed solutions to it. Such as
in demodulation, the bit recovered from demodulation can be
analyzed on oscilloscope but it is cumbersome. Therefore, an
asynchronous bit recovery algorithm is employed with serial
communication for representation and verification.
Figure 1: Phase change of BPSK signal.
A Test bed is essential for rigorous testing of the system.

Implementation of digital modulation schemes have become
more and more attractive and simple with the used of digital
signal processors (DSP) and Field Programmable Gate Arrays
(FPGA). Recently, researcher are utilizing the processing
power of FPGA and DSP together to implement systems.
Testing of BPSK modem is very important as all the data
transmitted through the transmission channel depends on the
correctness of implmentation of baseband modualtion.
The remaining paper is organized as follow. Section II
provides the previous work. Section III explains the proposed
methodology. Section IV explains experimental results.
Section V provides conclution of this paper.
Keywords- DSP, BPSK, modulator , demodulator.
I.
INTRODUCTION
Digital modulation is utilized to convert digital symbols into

waveforms to make it compatible with the characteristics of
transmission channel [3]. In digital modulation, a baseband
signal is changed into band pass signal compatible for
transmission. Similarly, demodulation is utilized in receiving
end to recognize the detected data.
Binary Phase Shift Keying (BPSK) is employed in
transmission and reception of telemetry and telecommand
data in satellites. It is the simplest form of phase shift keying
(PSK). Specially, at low frequencies (in kilohertz range) the
modulation and demodulation is performed digitally.
Generally it consists of two phases one is inphase with the
transmitted carrier and other is out of phase with the
transmitted carrier as represented in Figure 1. The two
conditions are generally represented as logic 0 and logic 1, the
formula associated with it represented in eq-1 and eq-2 [4].
II.
PREVIOUS WORK
Many researchers have designed and developed BPSK

modulator and demodulator in the Past [2] [5] [7] [8]. The
testing with genralized equipments are usually quoted in many
implementation. Including some results are acquired in the
form of software simulations. In this paper a testbed based
testing mechanism is proposed. Although it is formed using
traditional testing equipment but incombination with custom
solution.
III.
PROPOSED METHODOLOGY
Testing of BPSK modem is proposed in this paper. The test

setup is composed of two sets of modems. One is the designed
BPSK modem to be utilized in telemetry and telecommand
module and other is for testing of designed BPSK modem.
For testing of BPSK modulator and demodulator designed on
FPGA to be utilized for telemetry and telecommand. A testing
171
Proceedings
scheme is devloped which consists of a BPSK modulator with

variable frequency and data rate. This modulator is designed
on Digital signal processor. Similarly an FPGA is utilized for
debuging of the BPSK demodulator under test. On the other
hand, for testing of FPGA based Modulator Network analyzer
is utilized. The block diagram of the proposed setup is shown
in Figure 2.
Designed FPGA
TestBed
Figure 3: Constellation Diagram of BPSK modulator

FPGA Based BPSK
Modulator
FPGA Based BPSK

Demodulator
Network Analyzer
Based BPSK
Demodulator
DSP and FPGA Based

BPSK Modulator and
Debugging
Figure 2: Block Diagram for BPSK Test Bed.
IV.
EXPERIMENTAL RESULTS
Testing of BPSK modulator is significantly important for

reliable operation of satellite. Testing of telemetry modulator
data is acquired through on-board computer on FPGA
according to ECSS standard, the data is then converted to nonreturn to zero (NRZ) format and modulated. The modulated
signal is then tested using signal analyzer. A signal analyzer
has built in BPSK demodulator. The demodulator demodulates
the data and show constellation diagram. Let us first illustrate
the ideal result for BPSK constellation diagram in Figure 3. It
represents the possible symbols that may be selected by a
given modulation scheme as points in the complex plane.
Measured constellation diagrams can be used to recognize the
type of interference and distortion in a signal. For analyzing
the spectrum and constellation diagram of BPSK modulated
signal, Rhodes and Schwartz signal analyzer was used.
Signal analyzer contains a BPSK demodulator, which can
demodulate the signal accurately on wide range of frequencies
and construct constellation diagram. Figure 4 shows the
constellation diagram obtained from testing our BPSK
modulator with signal analyzer. That shows constellation
points position is very similar with the ideal results.
Figure 4: Signal analyser results for constellation diagram
Table 1 shows that the magnitude error and phase error are
in allowable number for satellite communication. Owing to
which the signal to noise ratio (SNR) is in the acceptable
range for space communication.
BPSK modulated signal spectrum as analyzed by signal
analyzer is shown in figure 5. Spectrum shows that the
difference of frequency of two peaks close to the center is 2
kHz which authenticates the data rate of 2kbps .As the symbol
rate is equal to the bit rate. As shown in the equation eq.3 and
eq.4.
R(s) = R (b) / n
(3)
For BPSK, (n = 1)
R(s) = R (b)
(4)
where R(s) is the symbol rate, R (b) is the bit rate and n
represent the number of phase changes. In case of BPSK
modulator the number of phase changes is 1. So symbol rate is
equal to bit rate.
172
TABLE I
PERFORMANCE CHARACTERISTICS OF BPSK MODULATOR
S.No
1
2
3
4
5
6
7
8
9
Modulation Accuracy
Category
Results
EVM
0.662
Magnitude Error
0.561
Phase Error
0.20
Amplitude Drop
0.01
Origin Offset
-73.86
Gain Imbalance
0.00
Quadrature Error
0.00
SNR (MER)
43.58
RHO
0.999
Peak
1.665
1.661
0.82
-
Proceedings
and uart transmission. So the data from ADC is stored in

memory and then transmitted through uart transmission.
Unit
%
%
%
dB
dB
dB
deg
dB
-
Figure 6: CCS Graph results of NRZ converted data and BPSK

modulated signal.
VI.
Our experiment showed testbed setup for BPSK modem.

Test setup for testing of BPSK modem using signal analyzer
and DSP/ FPGA test setup was proposed. The problems
encountered are discussed.
Figure 5: Spectrum representation of BPSK modulator on signal

analyzer.
Testing of demodulator is performed with DSP based BPSK

modulator and FPGA based debugging equipment. Result of
simulation on Code Composer Studios (Software used for
programming TI DSP Kits) are shown in Figure 6, Upper half
of the figure shows input converted to NRZ signal and in the
lower half corresponds to modulated BPSK signal [1]. The
BPSK modulator implemented on Texas Instruments
TMS320C6713 DSP kit has variable frequency and data rate
configurations. Althought in this testing application data rate
and frequency is fixed. The data from BPSK modulator is
provided through test computer serially and it modulates the
data. The modulated data is then input to FPGA based BPSK
demodulator. After the demodulation the data is transfered to
FPGA based Debugger for viewing the retreived output.
V.
CONCLUSION
ACKNOWLEDGMENT
This work was supported by Space and Upper Atmosphere
Research
Commission
(SUPARCO).
Authors
also
acknowledge the efforts of Mr. Muhammad Faisal and Mr.
Razi Iqbal for continuous support and efforts for completion
of this research work.
REFERENCES
[1]
[2]
PROBLEMS AND SOLUTIONS DURING TESTING
Several problems are encountered during the testing of BPSK

modem implemented on FPGA. Firstly, The bit recovered
after demodulation are continous, in long data stream it will be
difficult to monitor it in osciloscope. Therefore, a uart based
debug ging is implemented on FPGA for validation of
demodulator performance. Secondly, Analog to digital
converter has been tested seperately but on integration with
FPGA and DSP based modualator the rescaling was required.
Lastly, there was a transmission rate missmatch between ADC
[3]
[4]
[5]
173
G.Susinderrajan,
S.Saran
Kumar,
T.Shankar,
K.Sreeram,
P.V.Ramakrishna, "Design of a Low Cost All-Digital PSK/PM Satellite
Telemetry Transceiver", IEEE Electron Device Letters, vol. 20, 569
571, Nov. 1999.
S. Ruque, I. Ruiz, E. Carrin, "Simulation and implementation of the
BPSK modulation on a FPGA Xilinx Spartan 3 xcs200-4ftp256, using
Simulink and the System Generator blockset for DSP/FPGA", IEEE
Electron Device Letters, vol. 20,569571, Nov. 1999.
B.Sklar, DigitalCommunications Fundamentals and Applications, 2nd
ed. Englewood Cliffs, NJ: Prentice-Hall PTR, 2001.
S. Johanna, Ruque, D.Ruiz, Carlos, Simulation and implementation of
BPSK system on a FPGA board using system generator blockset for
DSP/FPGA, School of Electronics and Communication, Technical
university Loja 2005.
C.J. Harsha, D.H. Sandeep, A.S. Mali,"Hardware Implementation of
BPSK system on Virtex2-Pro FPGA using Xilinx System
Generatror".International Refereed Journal of Engineering and Science
(IRJES), vol. 2, pp 18-24, January 2013.
[6]
[7]
J.G. Proakis, Digital Communication, 5th edition, McGraw Hill New

York, 2001.
H. Malik, D.R. Rotake, M. Mahajan,"Design and Implementation of
BPSK Modulator and Demodulator using Vhdl". IOSR Journal of
Electronics and communication Engineering (IOSR-JECE), Vol. 9, pp
98-105, May-Jun 2014.
174
[8]
Proceedings
N. P. Shrirao, A.P. Thakare,"Design of Digital Modulators:BASK,

BPSK and BFSK using VHDL". International Journal of Advanced
Research in Computer Science and Software Engineering (IJARCSSE),
Vol.3 january 2013.
Proceedings
Design of a Fuzzy Logic Water Level

Controller
Taimoor Zahid
Nosheen Zafar
Institute of Space Technology,

Islamabad, Pakistan
taimoor187@yahoo.com
COMSATS Institute of Information Technology,

Islamabad, Pakistan
nosh_zaf@hotmail.com
Abstract This paper analyzes the effectiveness of

water level control using the concepts of Fuzzy Logic.
We have designed a Fuzzy Logic Controller for Water
Tank Level Control by using the linear control valve,
the level transmitter and the fuzzy control rules built in
the Fuzzy Logic Tool-Box and Simulink in MATLAB
and have then compared the control effect with that of a
Proportionalintegralderivative (PID). This design can
be very valuable for the industries where water level
control is extensively used and a slight deviation can
lead to major accidents and huge losses in revenue.
Therefore, it has become necessary to develop an
accurate and cost effective control system. The purpose
of such controller is to sustain a set point at a particular
value and be able to dynamically accept the new values.
Sometimes steady-state error and overshoot can no
longer be taken into account by the PID Controller due
to increase in the complexity, hysteresis, nonlinearity
and coupling of control object. While on the other hand,
with the development of the Intelligent Control Theory,
the application of Fuzzy Controller has become very
popular in practical industrial automation applications
and these problems can be successfully dealt with using
Fuzzy Control.
Control gives more attention to various parameters,

such as the time of response, the error of steadying
and overshoots [3].
This paper shows the development and
implementation of Fuzzy Logic Controller for Flow
Control Application because Fuzzy Controller can
achieve greater control results where limit of
overshoot is strict [4]. We have taken the level of the
water tank as an object and designed a Fuzzy
Controller in MATLAB. The Simulink models are
generated to see the working of tank. The controllers
are then added in those models to check and
understand the working of the designed controllers.
The control effect is examined and compared with
the effect of PID Controller. After designing the PID
and Fuzzy Controllers, the results are compared. This
approach of simulation by MATLAB allows us to
solve more complicated problems and develop the
measuring states to improve modeling and analysis of
system and to allow more control in the designs [5].
The overall goal of this paper is to show which
controller is better for flow and control applications.
Keywords Water Level Control, Fuzzy Logic, PID,

Tank, Mamdani.
I.
II.
INTRODUCTION
THE TANK PROCESS
Input, Process, Feedback and Output are the four

steps of tank process.
During the past several years, Fuzzy Control has

emerged as one of the most active and fruitful areas
of research in the application of Fuzzy Set Theory
[1]. Due to its robust, precise and orderly approach
for operating the controller, Fuzzy Logic can be used
in several applications. The field of Control System
Engineering has benefited the most from the success
of Fuzzy Set Theory. The reason behind this
unexpected success is high performance rate and low
cost. Although PID-Controllers are easy to install but
their performance can deteriorate quite fast when
used in non-linear systems [2]. Moreover, the
selection of the parameters of PID is also a difficult
task. If one does not choose the fine parameters, it
will badly impact the working of controller. Fuzzy
1) Input: A square-shaped waveform of

amplitude 0.5 and frequency 0.1 Hz is
generated from the signal generator as given
by (1).
y(t) = Amp * waveform(freq, t)
(1)
2) Process: The Tank-Process includes valve,

tank and controller. The water is pumped
into the tank via valve with controllable
degree of opening. Water flows out of the
tank through a hole in the bottom with area
0.05 m2. The height of the tank is 2m, and
the cross-sectional area of the tank is 1m2.
175
Technically, the water level in the tank can

be described mathematically using (2).
=
Proceedings
(b) Comparison: This block generates a plot and

that plot gives the comparison between the
reference signal and system output.
(c) S-function: This block provides a direct
access to the additional parameters to be
passed to S-function. The functions
parameters can be stated as variables
separated by commas or as MATLAB
expression. This function displays the
number of I/O ports plus the names of
specific S-functions. In our case, the built-in
file named tankdemo is used for
generating the animation. The S-function
used in this application refers to the
animtank which is built-in in the
MATLAB directory. ANIMTANK stands
for Animation of water tank system.
(2)
Where h is the water level [m],

is the influent
flow rate [m3/h] and
and
are parameters
depending on the tank and outflow areas. The
influent flow rate depends on the degree of valve
opening. In simulations, the water level in the tank is
controlled towards some pre-specified value
(reference signal) by letting the controller manipulate
the degree of valve opening. Eq. (2) will be used in
designing the Simulink model for controlling the
water level for the designed tank.
III.
A.
PID-CONTROL OF TANK WATER

LEVEL
PID Controller
PID controller refers to a Proportionalintegral

derivative controller. In order to overcome the
shortcoming of self-tuning PID controllers, fuzzy
logic and neural networks are utilized [2]. The
conventional PID is commonly utilized in controlling
the level, but the parameter of those controllers must
be turned by tuning method either in time response or
frequency response to meet their required
performances [6, 7]. The block diagram of PID
Controller with constants is shown in Figure. 2. P
element: e (t) is the proportional to the error at the
instant t, which is the present error. I element:
dt is proportional to the integral of the error
up to the instant t, which can be interpreted as the

accumulation as the past error. D element:
is
proportional to the derivative of the error at the
instant t, which can be interpreted as the prediction of
the future error. [8]
Figure.1 Water Tank System
3) Feedback: The signal is generated from the

signal generator and is further processed
through controller. This output from the
process becomes a feedback. This feedback
goes to a summer named error. It is named
so because it calculates the difference
between actual and desired water level. The
actual value is the one which is coming
through feedback and the desired is the one
which is coming straight through the signal
generator.
4) Output: The output can be generated via:
(a) To workspace: It writes the input to the
specified array or structure in MATLABs
main workspace. The To Workspace
block takes the signal as input and writes the
data of the signal into the workspace of
MATLAB. Data is not available until the
simulation is paused or stopped. This block
writes the data to an internal buffer during
the simulation and that data is written to the
workspace when a simulation is completed
or paused.
Figure.2 Block diagram of PID Controller
176
Proceedings
2) Valve: The water is pumped into the tank via

valve with controllable degree of opening. 0
means the valve is closed and 1 means that the
valve is fully open. The more the valve is open
(the closer the value is to one), the larger the
inflow to the tank. As illustrated in Figure. 5,
the product of the condition provided in the
limited integrator and inflow rate of the water
gives the rate at which water flows out. One of
the important parts of the valve is limited
integrator. The reason behind using the limited
integrator is to stop the output from surpassing
specific levels. The function of this block is
that, as soon as the output surpasses the limits,
it turns off the integral action for preventing the
integral wind up. The limits can be changed for
the period of simulation.
B. Simulink Model of Water Level Controller

Considering the tank process described above, the
Simulink Model is designed the as shown in the
Figure.4 and is saved with the name of
tankpid.mdl. Figure.4 illustrates that the process
includes PID Controller, valve and water tank.
1) Controller: The PID controller is the built-in
block in the Simulink. Its block diagram
representation is shown in the Figure.3. As
described earlier, it has three parts namely
proportional, integral and derivative. The input
is an error signal (difference between the
reference signal and system output) which when
comes into the controller is divided into three
respective parts and output from each of these
parts is then summed up to give a stable output.
The output of this block is the weighted sum of
the input signal, the integral of the input signal
and the derivative of the input signal [9]. In this
case, this controller is acting as a PD-controller
as the parameter which set for the integrator is
zero.
Figure.5 Simulink Model for Valve

Figure.3 Block Diagram for PID-Controller
Figure.4 Simulink Model for PID- Controller
3) Tank: The water level in the tank can be

described mathematically using (2) and tank
thus designed is shown Figure 6.
177
Proceedings
Controller in the Simulink model used in the next

sections of this paper.
IV.
FUZZY LOGIC CONTROL OF THE

WATER LEVEL TANK
The Fuzzy Logic Controller is developed through

following two approaches:
(a) Adaptive Neuro-Fuzzy Inference System
(ANFIS) Approach: For a particular system,
a FLC can be designed in different manners.
In the ANFIS approach, the FLC is trained
to produce the response similar to that of the
PID controller and once the training reaches
the required error tolerance, the PID
controller will be totally replaced by the
FLC.
(b) Design of a Fuzzy Inference System (FIS):
FIS System is developed by using Mamdani
approach using Fuzzy Logic Toolbox in
MATLAB. This involves designing and
tuning of the membership functions, I/O
rules, and the de-fuzzification technique.
Figure.6 Simulink Model of Water Tank
C. Execution
After the completion of the Tank-Design Procedure,
the next step is Execution. Executing the command
tankpid in MATLAB opens a Simulink program
where PID control of the tank process is simulated.
The desired water level (reference signal) and the
actual water level are saved into the workspace of
MATLAB
in
the
variable
comp.
By executing the MATLAB command plot
(comp.time,comp.signals.values), the reference
signal and the actual water level are plotted against
simulated time.
D.
A) Fuzzy Controller Design

The Fuzzy Controller for the tank process can be
designed by using the graphical tool in MATLAB.
First of all, a Fuzzy Controller (a FIS) is constructed
which uses the level and rate as input. The output
of the controller is the change in valve opening. The
general working of tank lies in its overflow and
underflow conditions. Taking care of these
circumstances, a brief description of the number and
shapes of input and output membership functions and
rule base is given in the following steps:
Results
The results obtained from the Simulink model of PID

Controller are shown in the Figure.7.
1) The Fuzzy Mechanism: As shown in the Figure.8,

the system has two inputs and one output. The
linguistic term used for the input is level and
rate and for the output is valve.
Figure.7 Output of PID- Controller
These results show that the level of water reaches to

the wanted output after passing some overshoots and
oscillations. The goal behind developing this water
level control to illustrate that the Fuzzy Logic
Controller can be used to control the water level and
to compare its control performance to the
conventional PID Controller. For this purpose, we
have replaced the PID controller with Fuzzy
Figure.8 Block diagram of Water Tank
(a) Input 1: For this FLC, the level (control

error) is defined as input of this Fuzzy
Inference System (FIS).This input basically
defines the control error for the water level.
178
The input range is [-1, 1]. It consists of three

membership functions namely high, good
and low. Trimf is used for good, and
Trapmf for high and low as a nominal
function. These membership functions are
shown in Figure.9.
Proceedings
(c) Output: The designed system has only one output

given by the linguistic variable valve. It basically
represents the position of the valve by accounting for
the change in valve opening. The output space range
is [-1 1]. The membership function for the output is
shown in Figure.11.
Figure.11 Membership function for output variable valve
Figure.9 Membership function for input variable level
2) Rule Base: The input-output relationships

are used to develop the IF-THEN rules for
the FIS [10]. From the two input variables
and one output variable total of four
following rules are developed.
1. If (level is low) then (valve is open_fast)
2. If (level is high) then (valve is close_fast)
(b) Input 2: The rate (flow rate) is defined as the

second input. Its range is [-1, 1]. The
membership functions are named as rising
and falling. Trapmf is used as a nominal
function.
Figure.10 Membership function input variable rate
Figure.12 Simulink Model of Fuzzy Controller of Water Tank
3. If (level is good) and (rate is rising) then (valve is

close_slow)
4. If (level is good) and (rate is falling) then (valve is
open_slow)
B. Design of the Simulink Model
179
Proceedings
[3] Qi Li, Yanjun Fang, Jizhong Song, Ji Wang, "The Application

of Fuzzy Control in Liquid Level System", 2010 International
Conference on Measuring Technology and Mechatronics
Automation, 03/2010.
After setting the parameters of the FIS membership

function and making the fuzzy controller; the FIS file
named tank.fis is exported to the workspace. The
reason behind the fact that the derivative of the
process variable is taken instead of the derivative of
error is that the derivative mode amplifies the sudden
changes in controller input signal and cause large
variation in the controller output. [11]. Now the tank
with Fuzzy Controller can be simulated by using
model shown in Figure.12 and the output comes out
to be as shown in Figure.13.
[4] Wangnipparnto S. and Tunyasrirut S., "Level Control in

Horizontal Tank by Fuzzy Logic Controller", 2006 SICE-ICASE
International Joint Conference, 10/2006.
[5] Duane Hanselman and Bruce Littlefield, "Book: Mastering
Matlab 5", ISBN# 0-13-858366-8, Prentice-Hall, 1998.
[6] W. K. Ho, C.C. Hang, and J. H .Zhou, "Performance And Gain
And Phase Margins Of Well-Known Pituning Formulas,"
Accepted For Publication In IEEE Trans. Contr. Syst. Techno,
1995.
[7] J. G .Ziegler and N. B. Nichols, "Optimum Settings for
Automatic Controller," By ASME Trans. Vol. 64, Pp.759-768,
1942.
[8] Aravind, Sekhar R, and B R Vinod, "A generalized method for
improving the performance of controllers by combining PID and
DELTA rule", 2012 Annual IEEE India Conference (INDICON),
2012.
[9] MATLAB documentation, Mathworks
(www.mathworks.com )
[10] Pathmanathan E, and R Ibrahim., "Development and
implementation of Fuzzy Logic Controller for Flow Control
Application", 2010 International Conference on Intelligent and
Advanced Systems (ICIAS), 2010.
Figure.13 Output of Fuzzy Controller
V.
CONCLUSION
[11] Silviu Ionita and Emil Sofron, The Fuzzy Model for Aircraft
Landing Control, Prentice Hall, 2001
In plant process control, the measurement and control

of flow are two very important factors. But these two
factors are susceptible to disturbances because of the
different variables affecting the flow rate. Therefore
for optimizing the performance of a plant, the flow
controllers are required to be robust. Fuzzy Logic
Controller is proposed to replace the conventional
PID controller due to its superior applicability and
robustness [12]. The concept of Fuzzy Logic is a
representation of the human thinking and decision
making process. The experimental results show that
Fuzzy Control is better than PID Control not only
because of simplicity of its implementation, better
control performance, robustness and overall stability
but also because it gives a reasonable consideration
to variety of parameters like overshoot and time of
response etc and for this reason, it tunes and develops
the FIS more intuitively as compared to the PID
Controller.
[12] Ronald R. Yager and Dimitar P. Filev, Essentials Of Fuzzy

Modeling And Control, Johan Wiley & Sons, 2002.
[13] D. Wu, F. Karray, I. Song, Water level control by Fuzzy
Logic and Neural Networks, IEEE Conference on Control
Applications, pp.3134-39, 2005.
[14] Namrata Dey, Ria Mandal and M. Monica Subashini, Design
and Implementation of a Water Level Controller using Fuzzy
Logic, International Journal of Engineering and Technology
(IJET), Jun-Jul 2013.
[15] J.
Yen, "Fuzzy logic-a modern perspective", IEEE
Transactions on Knowledge and Data Engineering, 1999
REFERENCES
[1] L.A. Zadeh, Fuzzy Sets, Informal Control, Vol. 8, Pp. 338353, 1965.
[2] Y. F. Chan, W. Chan and H. T. Mok "Adaptive Neurofuzzy
Network
Based
PI
Controllers
with
Multi-objective
Functions",Next-Generation Applied Intelligence, Vol. 5579, Pp.
604-613, 2009.
180
Proceedings
The Use of Radioactive Isotopes and Nuclear

Reactors in Space Application: Propulsion and
Power concept
Gohar Ali
Hira High school and Science College Shahdherai
Swat, Khyberpakhtunkhwa, Pakistan.
Email:105goharswati@gmail.com
Abstract
I.
Introduction
As we know that men have insatiable thirst in exploration

and discovery especially to unfold the mysteries of
universe. There are lots of frontiers that could be

covered in space exploration but unfortunately, there
are some reasons which affect it. Firstly, the
distances to be covered are extremely large, this
large distance in outer space makes it very difficult
to prospect for an astronaut involved in such a
mission. Even a simple mission to far planets like
Uranus, Neptune or even to nearest planet Venus
would take a long time. Hence in case of Venus, it
would need to take more than one year for such a
mission. However, to explore outer space of solar
system like to go to our nearest star PROXEMA
CENTAURI it would take more than millions of
years to reach there. Moreover, the second problem
is greater amount of energy required, which need for
an astronaut onboard for any outpost installation.
Luckily, the availability of nuclear technology allows
us to overcome these two problems. In this paper, it
is demonstrated that by using nuclear reactors or
Radioactive isotope decay like Americium-241 with
half life of 432.7 years or Uranium -235 using
Concept of propulsion and power can provide
enough energy to travel vast distances in space and
the energy need for an astronaut at ISS would be
easily available.
The huge distance in space becomes very difficult for a

space scientist and for astronaut to cover that. The
reason is the fuel need for such a mission not finds
easily because to goes out of solar system would takes
million of year.
Even
a simple mission in our solar system to the
terrestrial planet like to Venus or to Mercury would also

takes hundreds of years so, the one problem which
effects our mission the most is the propulsion need for a
space shuttle to covers such a distance,
And secondly we also know that energy requires for an
astronauts onboard for any outpost Installation also
becomes difficult.
Yeah it would be tempting to believe that all power in
space could be supplied by solar means since the sun is
available and free. However, in many cases the mission
some time takes place in the dark place and solar panel
is not always suitable for a mission (figure 1).
Keywords: nuclear propulsion, nuclear rocket, space power

application, nuclear power in space station.
181
Proceedings
Nuclear reactors have provided electrical power for
Nuclear reactors being a reliable and competitive source
some of the U.S. space programs greatest successes,
of energy they have some risk too and as well it also
including the Apollo lunar landings and the Viking
becomes difficult to do in some condition. The
Landers that searched for life on Mars. RTGs made
overcome of these problems are also demonstrated in
possible NASAs celebrated Voyager explorations of
this paper.
Jupiter, Saturn, Uranus and Neptune, RTG Power

sources are enabling the Galileo mission to Jupiter, the
international Ulysses mission studying the Suns Polar
Regions, and the Cassini mission to Saturn.
Pioneer 10 was launched from Cape Kennedy on 2
II.
Space power concept:
March 1972. It was the first interplanetary probe,

To remove the waste, to lift a load and as well as for
successfully navigating the asteroid belt before

Making rendezvous with Jupiter and Saturn The probe
was equipped with an array of instruments for
measuring such phenomena as the solar wind and the
magnetic and radiation fields surrounding Jupiter.
communication in International space station
or on
ground you will need sufficient power to do so for a

long time. Moreover it would be very difficult for to
generate electricity by thermal or chemical means in
microgravity environment [5].
However with the availability of radioactive isotopes

and as well as with the help of nuclear reactors the
Electricity and all the other requirements in ISS would
be easily available without any difficulty for several
years [5]
For the production of electricity heat energy can be
converted by many devices but the precise one is to use
Thermo electric generator which typically works on
Seebak Concept i.e
dV = SAB.(Kh - Kc)
Where
dV is Voltage difference,
Figure 1 : regimes of possible space power applicability
metals, & Kh - Kc
SAB is two dissimilar
is temperature gradient.
In this way Nuclear reactors can provide infinite power

for almost any time. However, they are not practicable
for applications below 15 kilo watt (kW).
182
Proceedings
Radioisotopes are considered to be best used for
position Or velocity (v), of space craft under the strong
continuous supply of low levels (up to 5 kW) of power
influence of gravity. [6]
or in combinations up to many times this value. For this

When a space craft moves it generally term is
reason, especially for long interplanetary missions, the
momentum (mv) so, the basic concept of propulsion is to
use of radioisotopes for communications and the
change the momentum (mv) of a space craft, and change
powering of experiments are preferred. The nuclear

process shown in Fig. 2 can either be a critical reactor or
in momentum is known as impulse (Imp). So, the basic

aim of propulsion in space is to create impulse, to
radioisotope fuel source such as plutonium oxide. In
measure the impulse is often term is specific impulse
either case the heat can be converted to electricity either
(Isp).[6]
statically through thermoelectric or a thermionic

converter, or dynamically using a turbine generator in
Specific impulse also known the exhaust velocity (ve)
one of several heat cycles (Rankin, Stirling, Brayton)..
(OR) the impulse per unit weight (
) but while
The nuclear workhorses for current space missions are

the RTGs and the TEGs powered by radioisotopes in the
discussing the engine in space there is no weight so it

can become impulse per unit mass
The difference arises here only by the acceleration due
to gravity (g).Where its value is
on earth.
However the value of g and reaction mass is not

important while the vehicle discussing in space. [4].
The two parameter specific impulse and the mass of the
rocket during launch and then in orbit is measure the
efficiency of propulsion energy.
Figure 2.The conversion of heat energy into kinetic or to
Where the equation for the measure of specific impulse
mechanical energy
as seen as (1)
(1)
Russian Federation that provide electricity through static
Where
(and therefore reliable) conversion at power levels of up
represent mass flow.
represent mass of propellant and
to half a kilowatt, or more by combining modules. And

Here it is seen that mass of propellant is inversely
that energy can be used for to run machinery. [2].
proportional to specific impulse, So, In chemical

propulsion there is used hydrogen and oxygen which
Space Propulsion concept
provides a specific impulse about 4400 m/s with mass

Nuclear reactor can also be used in rocket propulsion
ratio of earth escape of 15.[1]
system. Propulsion system in space is to change the
183
Proceedings
However americium OR hydrogen heated with fission
our solar system, because it provides a large specific
reactor can achieves twice the impulse with mass ratio
impulse as a result faster travel and the probability for
of 3.2. And with different core it can be generate much
complicated missions can be met, secondly nuclear
more impulse i.e. seven times having mass ratio of only
reactor has low molecular weight which can increase the
1.2 [1].
propulsive force per unit of propellant flow and allowing

us for increase the proportion of total weight of payload
III.
Nuclear rocket
[3].
Nuclear rocket is the rocket which use nuclear reactors

usually fission concept instead of chemical propulsion to
generate thrust it include nuclear thermal rocket (NTR),
IV.
Nuclear electric rocket (NER) or Hybrid (NTR/NER).
Microgravity effect on core

reactor
[2]
In microgravity environment every mass tends to be in
These rocket are propelled by the force of nuclear
state of motion (try to revolve from heavenly bodies
explosion, usually liquid hydrogen is heated via nuclear
near them due to centripetal force) So, The big challenge
fission which subsequently can be expanded in nozzle
of core reactors in microgravity environment is to
and thus accelerated at a high ejection velocity (6,000 to
control the flow of fission. It seems like very difficult
10,000 m/s) to generate thrust. There are also many
but under certain design it is possible to keep the
design are find for nuclear thermal propulsion i.e. liquid,
reaction self sustain. So the level of energy would be
solid or gas core rocket. [6]
maintained.
In core reactors usually the flow is
turbulent due to the randomness motion of molecule.

They use solid, liquid or gas core reactors respectively.
And hence it becomes difficult to keep flow within
Electro thermal propulsion rocket have been use in
specified boundaries. For this reason Reynolds number
many orbital mission in this type of rocket the propellant
flow are preferred for microgravity environment.
is heated electrically (heat energy of nuclear reactor are

converted into electricity) and then it accelerate the
ionized gas via electrostatic force at supersonic velocity
having range from 1,000 to 5,000 m/s and thrust range
0.01 to 0.5 N. But it has low Specific impulse therefore
it can be use only for interplanetary missions, however,
it can be use for long time instead of chemical rocket [2]
If it is not under the specified boundaries then the chain

reaction times would be increases due to uncontrolled
condition in the core reactor. And some parts of the core
reactor are exposed and hot spot would be appear on that
as well some parts of the core reactors would be damage
due to high temperature because the distribution of heat
is not possible in such circumstance.
The nuclear rocket has many advantages which

overcome on chemical rocket. Like it can be use in
much complicated mission even it can be travels to the
edge of our solar system or even to our nearest cluster,
nuclear space craft can be also used for to carry heavy
payload or to carry flyby, rover etc to the far planet in
184
Proceedings
radiation expose is measures in unit of Dose called

millirem. Over the course of year the average person
would expose to a total of 360 millirem to natural
radiation .In nuclear thermal rocket some of the
radioactive
element
would
be
release
into
the
atmosphere The particles which are hazardous to people

would remain high in the atmosphere for long course of
time gradually these particles being spread thinly across
the world. And eventually making their way on the
surface especially on ocean since these materials are
Insoluble once it reaches to the surface most of it would
Figure 3 Precise core reactors for microgravity environment
become trapped by the soil or by the ocean and not pose

a health hazard. Thus most of the release materials
1, Reflector moderator 2, Gaseous fissile zone 3,
wont be breath by people. But the small amount of
Working medium flow zone, 4, Fissile material
released material would be breath by the people and as
diminution replenishment 5, working medium Inlet.
well this small amount will be also distributed across the

world. So, In this way the amount to be breath will be
To overcome on this problem a special core is design
much less the person may receive it less than one
(shown in figure 3) to sustain fission reactor under
millirem in 50 years. This small radiation is negligible
control especially with this core the possibility of
compare to the 15,000 millirem a person will get it from
environmental hazard would be reduce. And various
natural resources over the period of 50 years. [7]
high temperature fuels are possible with this design. in

The fissile material for example Uranium or americium
would be located in the center of cavity enclosed by the
VI.
Result and conclusion
neutron moderator reflector. The working gas flow is

close to the cavity walls and it is heated by the high
In this paper we have emphasize on the importance of
temperature plasma radiation which causes the fission
nuclear reactors in space based application, where the
reaction to be self sustained under microgravity
nuclear reactors and radioactive isotopes would provides
conditions. [8]
infinite heat energy for long durations during journey of

the space craft and that energy can be converted into
V.
Radiation hazards of Nuclear

rocket
other form of energy and the need for power and

propulsion would be met.
But in microgravity
Expose of a person to radiation does not mean that the
environment it becomes difficult to do but however with
person would get cancer. People are exposing to
the help of gas core reactors and other special core
radiation on daily basis like on lesser extent from human
design would allow us to do so in microgravity
activities i.e.-rays. And radiation comes out from natural
environment too. And the space exploration would be
environment i.e. in the earth cosmic rays and radon. The
become more precious and peaceful.
185
VII.
Acknowledgment
I would like to thank national space agency SUPARCO

Space awareness society including Mr. Qamar Abbas
and Mr. Tahir hussain and My School principal Mr.
Sami ullah for to guide me as well for their support. I
also like to express my gratitude with SUPARCO
propulsion team for did my conceptual analysis.
I am also grateful to Mr. Rahman u din and Mr. Zulfiqar
Ali for Insatiable support and guide.
References
[1]. Rocket propulsion elements: introduction to
the engineering of rockets by George P. Sutton,
Oscar biblarz___7th edition.
[2].The role of nuclear power and nuclear
propulsion in peaceful exploration of space by
International atomic energy agency Austria
September 2005.
[3]. Basic of space flight by National Aeronautic
and space administration sec 2, 2002.
[4]. Advance space propulsion for the 21st Century
by Robert H .Frisbi a report from 9 August 2003 to
15 September 2003[5].Civilian usage of nuclear
reactors in space volume 1, science and global
security, 1989,
[6] Thruster precisely guide by Hiss, M martin and
K, Rachule October 2002.
[7]
Space
craft
power
(http://www.jpl.nasa.gov/cassini/)
for
cassini
[8] Fission fragments heating for space propulsion

by C. Rubbia
186
Proceedings
Proceedings
Impact of Thermal Aging on Microstructure and

Mechanical Properties of high Sn Content, Sn-Pb
Solders
Khurshid Alam2
Muhammad Aamir1, Riaz Muhammad1, Naseer

Ahmed1
Department of Mechanical and Industrial Engineering,

Sultan Qaboos University,
Muscat, Oman.

CECOS University of IT & Emerging Sciences,
Peshawar, Pakistan.
aamir@cecos.edu.pk
Abstract The microstructure and mechanical properties of
solder joints in electronic components are evolving when
exposed to thermal aging. The current work is based on the
investigation of high Sn content Sn-Pb solders i.e. 96Sn-04Pb.
Scanning Electron microscopy was carried out to observe the
microstructure examination followed by chemical composition
with elemental mapping using Electron dispersive X-ray.
Mechanical properties were examined using tensile tests. The
effect of thermal aging on microstructure and mechanical
properties has also been studied. The results show that
mechanical properties are highly influenced by its
microstructure after thermal aging. Due to aging effect, the
microstructure becomes coarsen and ultimately the mechanical
properties including yield strength and ultimate tensile
strength degrades but increase in ductility is obtained.
Keywords96Sn-04Pb,
Mechanical properties.
Microstructure
Solder joints are of great important in the electronic

industry, they are not only used to physically hold
assemblies together but also used for transmission of
electrical signals [8]. When electronic device is in use, the
solder connections faces mechanical stresses.
The on off switching of the system also leads cyclic
mechanical load and ultimately results in stresses in solder
joint [9, 10]. The decrease in the strength and microstructure
coarsening are also the result of high homologous
temperature and ultimately solder alloys will undergo creep
even at room temperature [11]. Pb based solders,
particularly 63Sn-37Pb or near eutectic Sn-Pb, has been
used in electronic industry for a long time. This is due to the
fact that Pb bearing solders have good metallurgical,
mechanical, solder ability, reliability, low cast, physical,
mechanical and metallurgical properties [12]. The use of Pb
in Sn-Pb solder is to reduce the surface tension which
ultimately results in improvement of wettability of the
solder joint [13, 14]. Sn is used as it is ductile in nature and
having good resistance to corrosion. High Sn concentration
also lead to high tensile and shear strength [15].
However, the concern about Pb toxicity and legislation to
ban on Pb bearing electronic products results in moving the
electronic industry towards environmental friendly green
products [16-18].
Therefore, electronic industries and academic institutes
accepted a challenge for characterizing new solders in terms
their mechanical properties, material prosperities and
manufacturing process compatibility with components and
printed circuit boards (PCB) [19].
The current work is based on developing and characterizing
a new solder with high Sn contents. The Pb content has been
limited upto 04% instead of 37% conventional Pb solders.
The reason for limiting Pb upto 04% is to get the properties
of Pb and make the environment as green as possible. The
microstructure including intermetallic particles and
mechanical properties have been examined at room
temperature and after thermal aging at 100C for 50 hours.
examination,
I. INTRODUCTION
The use of the solder alloys as metal joint can be dated back
to thousands of years and advancement in the electronics
industry are continuously in progress. However, the use of
surface mount technology (SMT) in printed circuit boards
(PCB) means that solders performs a leading role to make
electronic connections [1]. Surface mounting of the
components means that during impact loading the joint
transfer whole momentum of a component to the board.
Therefore, mechanical properties of the solder joint are of
great importance. Another important factor is the
accelerated aging due to relatively high temperature under
severe service condition. This also makes the joint in high
fraction of their melting point which results in
microstructural changes. Thermal aging is also responsible
for reduction in strength through grain growth after
microstructural coarsening process [2-6].
The mechanism for the decrease in strength in solder alloys
after aging is associated with coarsening of microstructure.
Coarser microstructure leads to fewer grain boundaries to
block the dislocation movements which ultimately cause
reduction in strength [7].
187
Proceedings
II. EXPERIMETAL WORK

Preparation of good sample is necessary for experimental
study for better characterization and mechanical properties.
Samples of 250g ingots of pure metals with composition
(wt. %): 96Sn-04Pb is prepared by putting them in the
crucible and heated in an oven for 30 minutes at room
temperature to get the molten metal. A die composed of two
parts of aluminum and a central steel plate have been used
to get dog-bone specimen after the alloy is then cooled in
air. The final dog bone shape specimen is shown in Fig. 1
[20].
In order to analyze the microstructure, the specimen is cut
into pieces and mounted in bakelite powder using mounting
press for proper handling and avoiding any distortion. The
temperature for using bakelite was is at 160 C for 09
minutes with 130 Kg/cm2. The mounted samples are
initially grinded with different grit sizes of silicon carbide
water proof sand papers. During grinding process, heat is
generated so tap paper is used for lubrication purpose to
avoid any damage to the surface. After sand paper grinding,
the specimens are polished with polycrystalline diamond
suspension with abrasive particle size of 6 m, 1 m and
0.25 m to give extra shine to the surface and then cleaned
with distilled water to remove any residue from the
specimen left during polishing. The specimens are then
etched for 45 seconds using 5% hydrochloric acid and 95%
ethanol. Proper care is necessary for good observation of
grain particles and grain boundaries for examination of
morphology under Scan electron microcopy. Before Scan
electron microscopy (SEM), samples are coated with silver
paste and placed in sputter coater and vacuum evaporator
for gold coating to make the material conductive. The SEM
images have been taken at different location and
magnification followed by Electron dispersive X-ray (EDX)
to confirm the chemical composition of the alloy.
For thermal aging, samples for microstructure examination
and dog bone specimen are exposed to 100C for 50 hours
in oven. The aging process is necessary to get the better
understanding of performance during extreme operating
conditions. For tensile testing, universal testing machine is
used to determine the mechanical properties. The data has
been taken at room temperature. The universal testing
machine with dog bone specimen is shown in Fig. 2.
Fig. 2 Universal testing machine with dog-bone shape specimen

A. Effect of Thermal Aging on Microstructure
The microstructure of 96Pb-04Pb alloy is described by low
and high magnification of images at different resolutions,
the images are taken for the as casted and after thermal
aging to examine the effect of changes in the morphology.
The as casted SEM images are shown in Fig. 3 and
thermally aged in Fig. 4. The chemical composition is
confirmed by Electron dispersive x-ray shown in Fig. 5. The
microstructure of 96Sn-04Pb is composed of soft Sn matrix
and hard IMCs of Pb. The black zone plate shape is the Sn
matrix and white color rods shaped are the IMCs of Pb. The
elemental mapping of Sn matrix and Pb particles are shown
in Fig. 6. These IMCs are usually brittle in nature in
comparison with soft Sn matrix and responsible for changes
in mechanical properties. The growth of these IMCs results
in coarsening of the alloy. The size of these IMCs depends
on many parameters including composition of alloy,
environmental conditions under sever temperature during
service and cooling rate. A highly temperature aged alloy
can produce coarse microstructure. By closely observing the
SEM images, it is observed that the Pb white particles
become wider and distributed in Sn matrix after thermal
aging. Since the properties of the solder joints are highly
microstructure dependent which changes reasonably during
its life time so, it is reasonably expected that after thermal
aging the mechanical properties would be worsen. Increase
in the growth of the grain size after thermal aging may also
leads in reducing the strength because larger grains create
fewer obstacles per unit area to the movement of
dislocation.
Fig. 1 Dog bone specimen [20]
188
Proceedings
Fig. 5 EDX Analysis of alloy
Fig. 3 SEM images at various magnification of as casted alloy
Fig. 6 elemental mapping of Sn matrix and Pb particles
B. Effect of Isothermal Aging on Mechanical Properties

As thermal aging affects the microstructure of the solder
alloy and ultimately influences mechanical properties
including yield strength and tensile strength, therefore the
impact of thermal aging over these properties is discussed in
this study. Yield strength, ultimate tensile strength and
elongation to failure are presented in Fig. 7 for the as
casted and thermally aged samples and summarized in
Table. I. Since the microstructure becomes coarse with
thermal aging, therefore, decrease in yield strength and
ultimate tensile strength is noted. It is investigated that 17.4
% reduction in yield strength and 25.5% reduction in
ultimate tensile strength have been occurred but 32.1%
increase in elongation is found. Elongation to failure is the
measure of ductility of material so a ductile material offers a
high elongation. It means that after thermal aging the
ductility of the alloy is increases. For larger grains, small
boundary area per unit volume exists therefore, a larger
grain size results in increasing the ductility.
Fig. 4 SEM images at various magnification of thermally aged alloy
189
[3].
[4].
[5].
[6].
Fig. 7 Mechanical properties of as casted and thermally aged alloy
[7].
TABLE I. TENSILE TEST DATA

Specimen Tensile Test
Description
Elongation @ Peak (mm)
Load @ Peak (N)
Strain @ Yield (%)
Strain @ Peak (%)
Stress @ Yield (N/mm2)
Stress @ Peak (N/mm2)
Condition
Alloy With
Alloy Without
Isothermal
Thermal
aging at 100C
Aging
up to 50 hours
1.0090
1.4870
435.00
324.00
1.2333
1.3946
1.6817
2.0231
16.856
13.919
19.595
14.595
[8].
[9].
[10].
IV. CONCLUSIONS
It is concluded that from the results and observations that
the properties of the solder joint are highly microstructure
dependent and a strong relationship exists between
microstructural and mechanical properties. It has been
observed that thermal aging leads to coarsen the
microstructure of the alloys and ultimately affects the
mechanical properties. Mechanical properties including
yield strength and ultimate tensile strength decreases but
considerable increase in ductility have been examined. This
also concluded that thermal aging is closely related to
mechanical properties. Therefore, great care must be taken
regarding microstructure evolution and mechanical
properties in designing and developing solders but for
electronics green and environment friendly Pb free solders
must be introduced.
[11].
[12].
[13].
[14].
[15].
REFERENCES
[1].
[2].
Prymark, J., et al., Fundamentals of microsystems

packaging. ed. RR Tummala, Mcgraw-Hill Book
Company, New York, USA, 2001: p. 420.
Jung, K. and H. Conrad, Microstructure coarsening
during static annealing of 60Sn40Pb solder joints: I
stereology. Journal of Electronic Materials, 2001.
30(10): p. 1294-1302.
[16].
[17].
190
Proceedings
Kang, J. and H. Conrad, Microstructure coarsening

during static annealing of 60Sn40Pb solder joints:
II Eutectic coarsening kinetics. Journal of
Electronic Materials, 2001. 30(10): p. 1303.
Jung, K. and H. Conrad, Microstructure coarsening
during static annealing of 60Sn40Pb solder joints:
III intermetallic compound growth kientics. Journal
of electronic materials, 2001. 30(10): p. 1308-1312.
Basaran, C. and Y. Wen, Coarsening in Bga solder
balls: modeling and experimental evaluation.
Journal of Electronic Packaging, 2003. 125(3): p.
426-430.
Tang, H. and C. Basaran, Influence of
microstructure coarsening on thermomechanical
fatigue behavior of Pb/Sn eutectic solder joints.
International Journal of Damage Mechanics, 2001.
10(3): p. 235-255.
Ma, H., et al. The influence of elevated temperature
aging on reliability of lead free solder joints. in
Electronic
Components
and
Technology
Conference, 2007. ECTC'07. Proceedings. 57th.
2007. IEEE.
Efzan, E. and A. Marini, A review of solder
evolution in electronic application. International
Journal of Engineering, 2012. 1(1): p. 2305-8269.
Frear, D., D. Grivas, and J. Morris, A
microstructural study of the thermal fatigue failures
of 60Sn-40Pb solder joints. Journal of Electronic
Materials, 1988. 17(2): p. 171-180.
Hacke, P., A. Sprecher, and H. Conrad,
Microstructure
coarsening
during
thermomechanical fatigue of Pb-Sn solder joints. Journal
of Electronic Materials, 1997. 26(7): p. 774-782.
Wassink, R.K., Soldering in Electronics, 1989.
Electrochemical Publication. 177.
Sharif, A. and Y. Chan, Dissolution kinetics of BGA
SnPb and SnAg solders with Cu substrates
during reflow. Materials Science and Engineering:
B, 2004. 106(2): p. 126-131.
Abtew, M. and G. Selvaduray, Lead-free solders in
microelectronics.
Materials
Science
and
Engineering: R: Reports, 2000. 27(5): p. 95-141.
Zeng, K. and K.-N. Tu, Six cases of reliability study
of Pb-free solder joints in electronic packaging
technology. Materials science and engineering: R:
reports, 2002. 38(2): p. 55-105.
Needleman, H.L., et al., The long-term effects of
exposure to low doses of lead in childhood: an 11year follow-up report. New England journal of
medicine, 1990. 322(2): p. 83-88.
Noor, E.E.M., et al., Wettability and strength of In
BiSn lead-free solder alloy on copper substrate.
Journal of Alloys and Compounds, 2010. 507(1): p.
290-296.
Abadin, H., et al., Toxicological profile for lead.
Atlanta (GA): Agency for Toxic Substances and
Disease Registry, 2007.
[18].
[19].
[20].
Harrison, M., J. Vincent, and H. Steen, Lead-free

reflow soldering for electronics assembly.
Soldering & Surface Mount Technology, 2001.
13(3): p. 21-38.
Smith III, E.B., Environmental Impacts and
Toxicity of Lead Free Solders. op. cit, 1999: p. 2.
Sadiq, M., R. Pesci, and M. Cherkaoui, Impact of
thermal aging on the microstructure evolution and
mechanical properties of lanthanum-doped TinSilver-Copper lead-free solders. Journal of
electronic materials, 2013. 42(3): p. 492-501.
191
Proceedings
Proceedings
Effect of aryl diazonium salt fucntionalization on the electrical properties of MWCNTs

and MWCNTs/CF reinforced polymer composite
Madni Shifa, Fawad Tariq, Kiran Ali, Rasheed Ahmed Baloch
Quality Management Directorate General
Pakistan Space and Upper Atmosphere Research Commission (SUPARCO)
75270 Karachi, Pakistan
Email: madnishifa@yahoo.com
Abstract Multiwall carbon nanotubes (MWCNTs) are kind

of inert and hydrophobic material. These MWCNTs do not
disperse properly and tend to agglomerate in polymer matrices
due to high surface area and strong van der Waal forces
between themselves. Numerous studies have been conducted
in the last few years to resolve the issue of CNTs dispersion
and solubilization in polymer matrix through functionalization
process. However, most of the functionalization treatments
introduce defects and results in deterioration of electrical
properties. In this study, MWCNTs were covalently
functionalized by direct grafting of aryl diazonium salt
(BF4N2-C6H4-NO2) and effect of functionlization on electrical
behavior was assessed under applied electric field in an
electrolytic cell. The MWCNTs were dispersed ultrasonically
in 2-propanol which act as an electrolyte in electrolytic cell.
Voltage was applied on copper electrodes and movement of
MWCNTs inside the electrolyte was carefully observed in-situ
under optical microscope. In case of functionalized MWCNTs
(f-MWCNTs), no effect was observed under applied voltage
up to 100 V whereas un-functionalized MWCNTs (pMWCNTs) tend to align themselves in regular fashion
showing conducting behavior. SEM analysis was performed to
see the consequence of functionalization treatment on
morphology of MWCNTs and optical microscopy was carried
out to see the effect of functionalization on the dispersion of
MWCNTs in cured and uncured state of polymer matrix.
MWCNT/CF reinforced polymer composites were fabricated
using hand layup and compression molding technique. DC
electrical conductivity & electromagnetic interference
shielding effectiveness (EMI SE) testing were performed to
examine the effect of MWCNTs functionalization on
fabricated samples. Test results showed that the
functionalization degraded the electrical properties of
MWCNTs.
Keywords Aryl diazonium salt, Functionalization,
Morphology of MWCNTs, SEM analysis, EMI SE
I.
INTRODUCTION
Carbon nanotubes (CNTs) are carbon based advance nanomaterials with unique set of structure and properties.
Commonly two types of CNTs are being used i.e. single wall
carbon nanotubes (SWCNTs) and multi wall carbon nanotubes
(MWCNTs). Outer diameter of SWCNTs is around 1-2 nm,
while the outer shell diameter of MWCNTs ranges from 10-
192
100 nm. A typical length range of CNTs varies from 10-1000

m and have aspect ratio of about 1000:1 [1]. CNTs are one of
the most promising materials known to men; their mechanical,
thermal and electrical properties are superb in all aspects [2].
Considering electrical properties of CNTs, electrical resistivity
is 0.4-0.7 m, maximum current density 107-109 Acm-2 and
the band gap of CNTs is 0-0.7 eV depending upon their
nature. Owing to exceptional electrical properties in
conjunction with high aspect ratio and low density (one sixth
time lower than steel), CNTs have gained great interest for
developing lightweight electrically conductive polymer
nanocomposite [3, 4].
CNT/epoxy polymer composites have been widely used in
advance military aircrafts, fairings, missiles, personal
communications and electronic packaging in which CNTs act
as nano-fillers in epoxy matrix [5].High electrical conductivity
and nanometer size of CNTs imply very low level of
percolation threshold in CNT/epoxy nanocomposite [6].
Exceptional electrical properties of CNTs directly related to
high aspect ratio and defect free structure of CNTs. Advance
applications of epoxy polymer composite demands high
electrical conductivity, electromagnetic interference (EMI)
shielding and electrostatic dissipation (ESD) characteristics.
Electromagnetic waves produced from some electronic
component decrease the performance of nearby electronic
system by introducing noise, disturbance, data loss, etc.
Hence, there is an intense need to shield the sensitive
electronic circuits form undesired signals through EMI
shielding material. One basic criteria of EMI shielding
material is that it should be electrically conductive.
CNT/epoxy composite is found to be promising candidate in
this regard [4, 7-12].
However, one key issue of CNT based polymer
nanocomposite is the dispersion of CNTs in polymer matrix.
Dispersion includes separation of CNTs bundles and then
stabilization of CNTs in polymer matrix. As CNTs
agglomerates into bundles due to high surface area and these
bundles exhibits inferior properties as compare to individual
CNT. Another key challenge is the good interfacial bonding
between CNTs and polymer matrix. As CNTs are inert and
atomically smooth material so there is a lack of bonding
between polymer matrix and CNTs [13-15]. Best route to
overcome stated problems is to attach functional groups
through surface modification treatments of the CNTs. There
are numerous approaches to functionalize CNTs depending
upon their end use and the choice of functional group. Among
them, modification of CNTs through aryl dizonium salt is

more versatile and easy with variety of functional groups.
However, fucntionalization process induces defects and
decreases the aspect ratio of CNTs thereby degrading the
electrical properties like conductivity, EMI shielding and ESD
of CNT reinforced polymer composite [13, 16-18].
Most of the literature discusses the effect of acidic
functionalization on the electrical properties of CNT based
polymer composite. However, the effect of functionalization
treatment through aryl diazonium salt on electrical properties
of MWCNTs and MWCNT/CF reinforced composite has been
seldom reported. Therefore a detailed study has been
conducted in which MWCNTs were functionalized through
aryl diazonium salt to attach covalent functional group. Effect
of applied electric field on the electrical behavior of fMWCNT and p-MWCNTs in electrolyte has been in-situ
observed under light microscope and discussed here in detail.
Moreover, electrical conductivity and EMI shielding
effectiveness of f-MWCNTs/CF reinforced polymer
composite and p-MWCNTs/CF reinforced polymer composite
was also evaluated and compared.
II.
Proceedings
Fig. 1. Schematic illustration of MWCNTs fucntionalization with 4Nitrobenzenediazonium tetrafluoroborate.
MATERIALS AND METHOD
A. Materials
MWCNTs used in this study were purchased from Cheap
Tube Inc, US and were synthesized through CCVD (Catalytic
Chemical Vapor Deposition). Specification of MWCNTs
given by manufacturer: outer diameter 10-20 nm, length 10-30
m, purity level > 95%, ash content <1.5%, specific surface
area 180-230 m2/g and bulk density 0.22 g/cm3. Specifications
details of carbon woven fabric which was used as
reinforcement in epoxy matrix are: areal density 3255,
thickness 0.50.005 mm, breaking strength > 270Kg/5cm in
wrap direction, weave style satin 3, 3K tow and equal no of
threads in weft and warp direction. Thermoset epoxy resin
(Bisphenol-A) containing a reactive diluent was chosen as
matrix because of its long pot life, low viscosity and
multifunctional properties. Cycloaliphatic amine was selected
as hardener and used in the ratio of 10:3.5 by weight with
epoxy resin.
B. Functionalization of MWCNTs
Functionalization of MWCNTs was carried out through a
facile method in which 5-10 mg MWCNTs were dispersed in
acetronitrile through bath sonication. After 30 minutes of
sonication, 0.15-5 equivalent of diazonium salt (BF4N2-C6H4NO2) was added and stirring of mixture was carried out for 5
days at room temperature. Then whole mixture was filtered
through 0.45 micron size PTFE filter paper and washed
several time with acetronitrile to remove unreacted dizonum
salt. Functionalized MWCNTs were dried at room temperature
under vacuum for 24 h before use [19]. Functionalization of
MWCNTs was visually confirmed by dispersing 5 mg of
pristine and functionalized MWCNTs in water in different
tubes and kept for 24 h. After 24 h pristine MWCNTs were
settle down at the bottom of tube whereas the functionalized
193
Fig. 2. Suspension of F (f-MWCNTs) and P (p-MWCNTs) in water after 24 h.
MWCNTs were remained suspended which confirmed the

functionlization of MWCNTs as shown in Fig. 2 whereas Fig.
1 demonstrates fucntionalization mechanism of MWCNTs.
C. Fabrication of MWCNTs/CF polymer composite
Functionalized MWCNT (0.2 wt% of matrix) was added to
ethanol and sonicated in sonication bath for 30 min to open up
MWCNTs bundles. Then epoxy resin was added to the fMWCNTs contained ethanol and further sonicated for 30 min
at 80C to evaporate ethanol and disperse MWCNTs in epoxy
resin. Upon cooling of MWCNT filled epoxy resin, hardener
was added and solution was magnetically stirred for 10 min to
thoroughly mix the hardener in the solution. Finally,
Proceedings
Fig. 3. Schematic of MWCNTs dispersion in matrix followed by hand layup on carbon fabric & curing of sample.
Fig. 4. MWCNTs/CF reinforced polymer composite fabrication through compression molding technique.
composite was fabricated through hand layup and compression

molding technique as shown in Fig. 3.
Aluminum die was in-house prepared for fabrication of
composite laminates through compression molding method
(Fig. 3a). Three layers of carbon fabric were cut according to
the dimension (75 mm width & 150 mm length) and placed in
the cavity of lower part (female part) of aluminum die one by
one and impregnated with MWCNT filled epoxy resin through
applicator brush. Upper part (male part) of the die was placed
on the lower part and aligned through guided pins. Sufficient
pressure was applied by hand tightening of guided pins so that
excess resin escaped out through drain holes. Total 3 layers of
carbon fabric along with MWCNT filled epoxy were used to
get final 2 mm thick composite. Composite sample was cured
by placing the aluminum die in an oven at 80C and 120C for
30 min and at 160C for 2 h. After curing process, sample was
removed from the die and trimmed to get smooth edges (Fig.
4b & 4c). Similar process was also carried out to fabricate pMWCNTs/CF reinforced polymer composite samples.
III.
conductive. Glass slide was observed under FE-SEM at high

magnification. Similar process was also employed to examine
the p-MWCNTs morphology under FE-SEM.
B. In-situ observation of MWCNTs under applied electric
field
Functionalized MWCNTs were dispersed in 2-propanol
through sonication process which act as an electrolyte in
electrolytic cell. Electrolytic cell was designed by connecting
two copper wires with plastic jar through holes (Fig. 5a). One
wire was connected to the positive terminal of battery and
second wire was connected to the negative terminal of the
battery which serves as anode and cathode respectively. DC
voltage of 20V, 40V, 60V, 80V and 100V was applied and the
movement of MWCNTs in electrolyte was observed under
light microscope (Optika, Italy) at 100x magnification as
shown Fig. 5b. Similar process was employed to assess the
effect of applied electric field on the behavior of p-MWCNTs
in electrolyte.
C. Optical microscopy
Optical microscopy was also carried out in order to see the
dispersion state of MWCNTs in epoxy matrix. A drop of fMWCNTs filled epoxy matrix was taken on slide, gently
pressed by placing another slide on it and observed under
optical microscope. Slides were then placed in oven and
allowed to cure at prescribed curing cycle. Slides were again
CHARATERIZATION TECHNIQUES
A. Field emission scanning electron microscopy (FE-SEM)

FE-SEM (MIRA 3 TESCAN) was carried out in order to
observe the effect of functionalization on the morphology of
MWCNTs. Functionalized MWCNTs were dispersed in
ethanol through bath sonication , and one drop was taken on
glass slide and gold coated to make the sample electrically
194
Proceedings
of the sample. Testing was performed on equally dimensional

samples (150 x 75 x 2 mm) of f-MWCNTS/CF reinforce
polymer composite and p-MWCNTs/CF reinforced polymer
composite.
IV.
A. FE-SEM
Figure 6a and 6b shows the high resolution SEM images of
f-MWCNTs and p-MWCNTs respectively. Opening of end
caps and shortening of length was evident in case of fMWCNTs as shown in Fig. 6(a) while Fig. 6(b) demonstrates
that p-MWCNTs have long length and damage-free structure.
High aspect ratio and defect free structure of nanotubes is
prime requirement for exhibiting good electrical conductivity.
Morphological study through FE-SEM confirms that the aryl
diazonium salt functionlization has destroyed the physical
structure and decreased the aspect ratio of nanotubes, which in
turn decreased the electrical conductivity of MWCNTs.
Fig. 5. Picture of (a) Electrolytic cell (b) In-situ observation of MWCNTs

under electric field through light microscope.
examined after curing under the microscope to witness the

effect of curing cycle on dispersion state. Similar process was
carried out with p-MWCNTs containing epoxy resin.
D. DC electrical conductivity measurements
DC electrical conductivity of fabricated samples (both fMWCNTs/CF and p-MWCNTs/CF) was measured at room
temperature in accordance with ASTM standard D4496-98
and compared with the results of neat CF reinforced polymer
composite. Standard four point probe method was used to
determine the surface and volume conductivity via four Kelvin
pins and LCR meter.
E. EMI shielding effectiveness (EMI SE) test
EMI SE test was carried out in Anechoic chamber using
calibrated Vector Network Analyzer (Rohde & Schwarz) over
the frequency range of 7-14 GHz. Attenuation was measured
by recording S21 parameter (return loss) between source and
receiving antenna inside the anechoic chamber without any
sample. Then test sample was placed between the antennas
(source & receiving) and attenuation was again measured by
noting down the value of S21 parameter. The difference
between S21 values of sample and without sample gives the
value of attenuation (i.e. EMI SE) caused by the sample.
Attenuation is described in terms of shielding effectiveness
(SE), measured in decibel (db) and shows the reflective nature
195
B. In-situ observation of MWCNTs under applied electric

field
When DC electric field was applied up to 80 V no
significant effect was observed on MWCNTs under
microscope. However at 100 V fraction of p-MEWCNTs
moved from cathode to anode and network of nanotubes is
formed between electrodes. Under applied electric field
MWCNTs experience force due to polarization effect and
move from cathode to anode get discharged and absorbed on
the anode. Then adsorbed MWCNTs become the source of
adsorption of further nanotubes because of high electric field
at the tip which is connected to anode. In this way, conductive
pathway of p-MWCNTs is formed from anode to cathode
showing the high conducting nature of p-MWCNTs as shown
in Fig. 7 (a). In case of f-MWCNTs most of the nanotubes
under applied electric field attached to the anode without
development of conductive pathway because of imbalance
polarization effect due to structural damage during
fucntionalization process. Similar effect was also observed
with acid functionlized MWCNTs under applied electric field
by other researchers [20]. Transfer of electric current is not
possible between the electrodes without formation of
conductive path suggesting the poor electrical conductivity of
f-MWCNTs. Arrows head show the absence of conductive
pathway in Fig. 7 (b).
C. Optical microscopy
Figure 8 (a, b) shows the optical images of 0.2 wt % fMWCNTs and p-MWCNTs filled epoxy matrix in uncured
and cured state. Functionalized MWCNTs were more
uniformly distributed in the matrix and less amount of submicron size aggregates were present as compared to pMWCNTs-epoxy matrix in uncured state. f-MWCNTs were
dispersed homogenously in matrix during sonication process
due to the bonding between epoxy and functional group
grafted on MWCNTs. However, in case of p-MWCNTs
homogenous distribution was difficult in matrix because of
Proceedings
Fig. 7. Optical images showing the comparison of conductive pathway

formation of (a) p-MWCNTs (b) f-MWCNTS
motion get seize during polymerization. However, fMWCNTs remained suspended in the matrix and very few
aggregates formed during curing stage because bonding
between epoxy and functional group resist re-agglomeration
and coagulation of f-MWCNTs.
D. Eelectrical conductivity
Conventional polymers are usually electrical insulators and
do not fulfill the requirements of EMI and ESD
characteristics. Polymers are made conductive generally by
addition of conductive fillers to protect against EM waves and
electrostatic charge (ESC). Fig. 9 shows the bar graph of
volume and surface electrical conductivity of neat CF
reinforced polymer composite, p-MWCNT/CF reinforced
polymer composite and f-MWCNT/CF reinforced polymer
composite. Conductivity of neat CF reinforced polymer
composite is due to the conductive nature of carbon fabric.
Figure 9 illustrates that the volume as well as surface
conductivity both significantly increased by the addition of 0.2
wt% of p-MWCNTs. This increment in conductivity was
attributed to high aspect ratio and formation of 3d conductive
network of MWCNTs inside the composite. However,
addition of 0.2 wt% f-MWCNTs has not increased the
conductivity of polymer composite up to appreciable extent. It
was concluded from the conductivity measurements that the
functionalization process has in fact adversely affected the
conductive nature of MWCNTs by shorting the length and
creating defects in the structure of MWCNTs which in turn
ruined the conductivity of composite sample.
Fig. 6. FE-SEM images of (a) f-MWCNTs and (b) p-MWCNTs.
high electrostatic forces present between MWCNTs.

Similarly, in cured state amount of aggregates in f-MWCNTs
was lowered than that of p-MWCNTs in epoxy matrix. Since
curing is a lengthy polymerization process, p-MWCNTs find
time to get re-entangled due to vander walls forces before their
196
E. EMI-SE
Electromagnetic radiations can be attenuated by three basic
mechanisms: reflection, absorption and multiple reflections.
Figure 10 exhibits the SE of f-MWCNTs/CF and pMWCNTs/CF reinforced polymer composites against EM
waves at different frequencies. EMI SE is sensitive to
Proceedings
interface between carbon fabric and the matrix through

multiple reflection phenomena. And some part of EM waves
was absorbed by MWCNTs but mainly EM waves were
attenuated by reflection mechanism. As shown in Fig. 10 that
the p-MWCNTs composite demonstrated overall better EMI
SE which is governed by high electrical conductivity of pMWCNTs. For reflection of EM waves, 3D conductive
pathway is required which is better provided by p-MWCNTs
than f-MWCNTs. Maximum SE was observed by p-MWCNTs
sample at frequency range of 7-9 GHz with a peak about 70 db
at a 7.75 GHz frequency. However, in the frequency range of
10-11.5 GHz f-MWCNTs composite performed similar way as
p-MWCNTs sample and maximum SE about 46 db was
observed at 11.4 GHz in case of f-MWCNTs sample.
Minimum attenuation of about 14.8 db was observed at 13.7
GHz by p-MWCNTs sample and in case of f-MWCNTs
minimum attenuation of 9.4 db was attained at the frequency
of 13.5 GHz. From the test results, it can be inferred that pMWCNTs composite performed better than f-MWCNTs in
shielding EM waves in the whole tested frequency range with
the maxima at around 70 db.
Fig. 8. Optical images showing the dispersion of (a) f-MWCNTs and (b) pMWCNTs in uncured and cured state of polymer matrix.
V.
CONCLUSION
Effect of functionalization by aryl diazonium salt on

electrical properties of MWCNTs was assessed through
applied electric field test in electrolytic cell, electrical
conductivity measurement and EMI SE test. Following
conclusions were drawn on the basis of test results:
SEM microscopy showed that the functionalization
process has damaged the physical structure of nanotubes.
It was found from the optical microscopy that the
dispersion of f-MWCNTs in the host matrix was
improved.
Amount of aggregates of MWCNTs increased upon
curing in both cases (p-MWCNTs & f-MWCNTs) but fMWCNTs showed more resistance to agglomerate
formation during curing stage.
It was deduced that high aspect ratio and defect-free
structure of MWCNTs is essential for formation of
conductive pathway under applied electric field. Pristine
MWCNTs aligned themselves in a regular fashion under
applied electric field whereas the f-MWCNTs do not
respond appreciably to the applied field.
DC electrical conductivity measurements showed that
the conductivity of f-MWCNTs was decreased as a result
of functionalization. The reason for deterioration in
electrical conductivity was attributed to the introduction
of defects in the MWCNT sidewalls and shortening of
length.
Pristine MWCNTs based composite sample outclassed
the f-MWCNTs composite in EMI SE testing over the
tested frequency range of 7-14 GHz.
3D conductive pathway formation of MWCNTs inside
the matrix is the prime requirement for exhibiting good
electrical properties and this in turn drastically affected
by the choice of functionalization procedure.
Fig. 9. Volume and surface dc electrical conductivity of composite samples.
frequency and dependent on the type of MWCNTs addition.

Absorption and multiple reflections play an important role in
shielding against EM waves but dominating mechanism at
lower frequency is reflection which is directly related to
electrical conductivity of the sample [21]. Layered structure of
composite is responsible for the lost of EM waves at the
197
Proceedings
Fig. 10. Comparison of shielding effectiveness of f-MWCNTs/CF and p-MWCNTs/CF reinforced polymer composite.
[9]
ACKNOWLEDGMENT
Authors would like to thank Mr. Ahmed Bilal (Chairman,
SUPARCO) for approval and provision of facilities. The
authors also gratefully acknowledge Mr. Wasim Nawaz for
performing EMI testing and Ms Noureen Owais for
providing assistance in electrical conductivity measurements
of the samples.
[10]
[11]
REFERENCES
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[12]
Z.L. Wang, W. H. Manu and Dechang Li, Structure and growth of CNT
films by Pyrolysis, Cheml Phys Lett, V. 316(5,6) , pp-349, 2000.
Ajayan PM, Zhou OZ, Applications of carbon nanotubes. Carbon
Nanotubes,. pp.391-425, 200.
R. Lee, P. Nikolaev, H. Dai, P. Petit, J. Robert, C. Xu, Y. Hee Lee, S. G.
Ki, A. G. Rinzler, D. T. Colbert, G. E. Scuseria, D. Tomanek, and R, E.
Smalley, Thesis, Crystalline Ropes of Mettaic Carbon Nanotubes,
Science Vol. 273. No. 5274, pp. 483-487, 1996.
Prabhakar R, Bandaru Electrical Properties and Applications of Carbon
Nanotube Structures, Reviews journal of Nanoscience and
Nanotechnology Vol.7, pp.129, 2007.
Christopher Kingston, Richard Zepp, Anthony Andrady, et al. ReviewRelease characteristics of selected carbon nanotube polymer
composites,CARBON Vol. 68 pp.3357, 2014.
Sandler JKW, Kirk JE, Kinloch IA, Shaffer MSP, Windle AH, Ultralowelectrical percolation threshold in carbon-nanotube-epoxy
composites. Polymer, 2003, Vol. 44, pp.5893-9, 2003.
Baughman R. H, Zakhidov A. A, de Heer, W. A, Carbon nanotubes-the
route toward applications, Science, Vol. 297, pp. 787-792, 2002. ISSN
0036-8075
Bauhofer, W., & Kovacs, J. Z, A review and analysis of electrical
percolation in carbon nanotube polymer composites, Compos. Sci.
Technol., Vol. 69, pp. 1486-1498,ISSN 0266- 3538
198
[13]
[14]
[15]
[16]
[17]
[18]
Singh BP, Choudhary V, Saini P, Mathur RB, Designing of epoxy

composites reinforced with carbon nanotubes grown carbon fiber fabric
for improved electromagnetic interference shielding, AIP Adv 2012;2:6
Mathur RB, Singh B.P, Tiwari P.K, Gupta T.K, Choudhary V
Enhancement in the thermomechanical properties of carbon fibre-carbon
nanotubes-epoxy hybrid composites, International Journal of
Nanotechnology, Vol. 9, pp.1040-1049, 2012.
Huang Y, Li N, Ma Y, Feng D, Li F, He X, et al, The influence of
single-walled carbon nanotube structure on the electromagnetic
interference shielding efficiency of its epoxy composites,
Carbon,Vol.45, pp.1614-21, 2007.
Li N, Huang Y, Du F, He XB, Lin X, Gao HJ, et al. Electromagnetic
interference (EMI) shielding of single- walled carbon nanotube epoxy
composites, Nano Letters, Vol.6, pp.1141-5, 2006.
Kannan Balasubramanian and Marko Burghard, Chemically
Functionalized Carbon Nanotubes reviews, Small, Vol. 1, No. 2, pp.
18092, 2005.
A.KANAPITSAS, C.TSONOS, Study of electrical / dielectric and
thermomechanical
properties
of
polymercarbon
nanotubes
nanocomposites, In Proceedings of the 1st WSEAS International
Conference on NANOTECHNOLOGY.
R.B. Mathu r, Shaila ja Pande, B.P. Sing h, T.L. Dhami, Electrical and
Mechanical Properties of Multi-Walled Carbon Nanotubes Reinforced
PMMA and PS Composites, POLYMER COMPOSITES-2008
C.VELASCO-SANTOS, A.L.MARTINEZ-HERNANDEZ and V. M.
CASTANO, Review Carbon nanotube-polymernanocomposites: The
role of interfaces Composite Interfaces, Vol. 11, No. 8-9, pp. 567586,
2005.
Najiba Abdullah Al-Hamdani, Preparation and Electrical Properties of
Epoxy Resin Reinforced with Functionalized Carbon Nanotubes, IOSR
Journal of Applied Physics (IOSR-JAP) Volume 6, Issue 4 Ver. I, pp.
54-56, Jul-Aug. 2014
Jeffrey L. Bahr, Jiping Yang, Dmitry V. Kosynkin, Michael J.
Bronikowski, Richard E. Smalley, and James
M.
Tour,
Functionalization of Carbon Nanotubes by Electrochemical Reduction of
Aryl Diazonium Salts: A Bucky Pap, J. Am. Chem. Soc,Vol. 123,

6536-6542, 2001.
[19] Anastasia A. Golosova, Christine M. Papadakis and Rainer Jordan
Chemical functionalization of carbon nanotubes with aryl diazonium
salts Mater. Res. Soc. Symp. Proc. Vol. 1362 2011 Materials
Research Society DOI: 10.1557/opl.2011.1141
[20] Abu Bakar Sulong, Nurhamidi Muhamad, Jaafar Sahari, Rizauddin
Ramli, Baba Md Deros, Joohyuk Park Electrical Conductivity
Behaviour
of
Chemical
Functionalized
MWCNTs
Epoxy
Nanocomposites European Journal of Scientific Research ISSN 1450216X Vol.29 No.1, pp.13-21, 2009.
[21] Simon
Rea,David
Linton,
Eddie
Orr,
Jonathan
McConnellElectromagnetic Shielding Properties of Carbon Fibre
Composites in Avionic Systems,Microwave Review, June, 2005
199
Proceedings
Proceedings
Design of C band Slotted Waveguide Antenna with

High Impedance Bandwidth and Improved
Reflection Coefficient for SAR applications
Sania Nazir1, Masab Iqbal2, A.K.M. Shafaat Ali3
Satellite Research and Development Center,
Space and Upper Atmosphere Research Commission, Karachi, Pakistan
sania_tl@yahoo.com, masabiqbal@gmail.com,shafaat_32@yahoo.co.uk
AbstractWith the advancement of technology, synthetic
aperture radar (SAR) has acquired great importance in the field
of remote sensing. For optimum performance of SAR systems,
selection and design of antenna choice is very critical. In certain
applications, specially for airborne purpose, SAR antenna should
be efficient, conformal, medium to high gain, capable of
providing wide beam in range while narrow beam in azimuth
direction and either mechanically or electronically steerable.
These performance parameters can be met with slotted
waveguide antennas and are an ideal choice for radar and
Airborne SAR applications. Slotted waveguide arrays can be
either standing wave (resonant) or travelling wave (nonresonant). In this paper, a slotted waveguide antenna which
resonates at C band is presented. This paper discusses the design
of a bottom fed linear resonant slotted wave guide antenna array
with longitudinal shunt slot elements. This design is optimized
using advanced EM simulation software in order to achieve
improved impedance bandwidth and good reflection coefficient.
This improvement is achieved by introducing non-uniform inter
element center to center spacing and slot center to short end
spacing as odd multiple of guide wavelength (g/4). Reflection
coefficient value at the desired frequency comes out to be < -35dB
with an impedance bandwidth of > 100 MHz. Measurement
results are in conformance with the simulated results with a gain
of approximately 20 dB.
Keywords: Slotted Waveguide,
impedance bandwidth.
I.
reflection
coefficient,
INTRODUCTION
In this advanced era of remote sensing and imaging, synthetic

aperture radar (SAR) has attained great significance due to its
all weather, day/night imaging capability. It is basically a
technique of generating the effect of long antenna with
physically small antenna. It consists of long array of
successive and coherent radar signals initially transmitted and
then received by physically small antenna every time when it
moves along a flight path. Thus antenna plays an important
role in the overall performance of SAR system. For SAR
applications, the antenna should have following
characteristics:
200
Efficiency: It is essential for SAR antenna that the received

echo signal must be strong enough to provide required signal
to noise ratio(SNR). For receiving echo signals with adequate
SNR the antenna efficiency must be higher.
High gain: In SAR systems, specially pulsed SAR systems
whether space borne or air borne, it is required to transmit
very high power short duration pulses to achieve high
effective isotropic radiated power(EIRP). These high power
pulses are generated by either TWTA or SSPA. If transmitter
is not capable of generating very high peak power then
required EIRP can be achieved by using high gain antenna.
Wide beam in Range and narrow beam in Azimuth: SAR is a
two dimension imaging system, along track direction
(azimuth) and cross track direction (range). The swath size
which is covered by the system is specified in terms of
minimum and maximum look angles in range direction. It can
be stated that range foot print defines the swath size to be
covered on ground. Hence the beam must be wide in the range
direction to meet the requirement of higher swath width.
Mechanical or electronic steering: SAR can be operated in
several imaging modes such as strip map, spot light and scan
SAR. Among these modes, strip map mode has fixed antenna
beam in azimuth as well as range direction. While for spot
light and scan SAR which provide high resolution and larger
swath width respectively antenna beam must be either
mechanical or electronic steerable to achieve high resolution
and swath width especially in space borne SAR systems.In
this paper, a resonant C band slotted waveguide antenna is
presented used for Strip Map Air borne SAR applications.
II.
ANTENNA DESIGN
The proposed antenna is a slotted waveguide antenna

working at C band. Operating frequency range is in between is
5.2 to 5.4 GHZ. This slotted waveguide antenna is resonant,
linear antenna array with longitudinal shunt slot elements cut
along the length of waveguide. Waveguide end is short
circuited which in turn converts waveguide into a resonant

standing wave structure. Thus the antenna beam is fixed and
antenna array is non- scanning. Resonant arrays can be either
end fed or center fed. Proposed antenna is end fed where one
end of waveguide is fed while other end is terminated in short.
The antenna configuration is shown in figure 1.
In this configuration the some of the most important
parameters which affect the performance of antenna are; (a)
number of slots which determine gain (b) slot length which
determines the resonant frequency (c) inter slot spacing and
slot location in the waveguide. For resonant arrays it is
recommended that the inter element center to center spacing
must be uniform, which is equal to half of guide wavelength
(g/2) and the distance from short wall of waveguide to the
center of last slot should be equal to quarter of guide
wavelength (g/4). Where g is the guide wavelength and is
given by Eq1.
g = / c2 -2
Proceedings
III.
SIMULATION RESULTS
The proposed antenna is modeled, simulated and analyzed in

Advance EM Simulation software. Isometric view of the
modeled structure is shown in figure 2. After optimization,
this antenna provides a gain of 20 dB. The simulated gain
plots are shown in figure 3 and 4.
(1)
Where,
c=2a= Cut off wavelength
a= Longest dimension of rectangular waveguide
= Operating wavelength
Figure 2 HFSS model of proposed antenna
This design is optimized for achieving higher impedance

bandwidth and reflection coefficient with non-uniform spacing
and last slot center to short end spacing as odd multiple of
quarter guide wavelength (g /4). Slots along the length of
waveguide are modeled in such a way that the center to center
slot spacing between consecutive slots is (g), while the center
to center inter element spacing between alternate slots is lesser
and higher than half of guide wavelength. Last slot center to
short end spacing is either 3g /4 or 5g /4 or higher instead of
g /4 in order to achieve higher reflection coefficient and
impedance bandwidth.
In the proposed design, non-uniform spacing between the last
slot center to short end spacing is 5g /4 and the length of
waveguide is 826 mm with 22 longitudinal shunt slot elements
for achieving gain of approximately 20 dB.
Figure 3 Simulated gain plot
Figure 1 Basic structure of resonant slotted waveguide antenna array with

shunt slot element
201
Figure 4 Simulated 3D radiation plot
Figure 5 Simulated Reflection coefficient (S11) plot
Proceedings
Figure 8 Measured Azimuth cut
The simulated reflection coefficient is less than -40 dB at

desired frequency while the impedance bandwidth is greater
than 100 MHz. Figure 5 shows the reflection coefficient plot.
IV.
MEASUREMENT RESULTS
The proposed antenna is fabricated on wire cut machine in

order to have the slots fabricated with precision along with
maintaining very strict inter element spacing. The fabricated
antenna is shown in figure 6. The measured reflection
coefficient plot is shown in figure 7. Measured reflection
coefficient at desired frequency is -43 dB while impedance
bandwidth is 145 MHz. Antenna radiation pattern measurement
was carried out in Anechoic Chamber Test facility. Measured
azimuth and elevation cuts are shown in figure 8 and 9 the
measured gain at our desired frequency is about 20 dB.
Figure 9 Measured Elevation Cut
V.
CONCLUSION
A resonant C band slotted waveguide antenna is designed,

fabricated and tested as per design requirement. Gain of the
antenna is approximately 20 dB with -43 dB reflection
coefficient value. Impedance bandwidth achieved is 145 MHz
for VSWR < 2.
VI.
ACKNOWLEDGMENT
First I would like to thank Almighty for all His blessings

and providing me the strength for the successful completion of
this task. I would also like to thank SUPARCO for providing
me the opportunity and necessary support and guidance from
the team members for accomplishing this task.
Figure 6 Fabricated slotted waveguide antenna array
VII. REFERENCES
[1] Antenna engineering Handbook Richard Clayton Johnson.

[2] Waveguide slot antenna array by Ronald A Gilbert BAE systems.
[3] W1 GHZ Microwave Antenna Book Paul Wade W1GHZ.
[4] Resonant length calculation and radiation pattern synthesis of
longitudinal slot antenna in rectangular waveguide By M.
Mondal, A. Chakrabarty Progress In Electromagnetics Research
Letters, Vol. 3, 187195, 2008.
[5] A Design of Slotted Waveguide Antenna Array Operated at X
band Kuo-Lun Hung, His Tseng Chou. Wireless Information
Technology and Systems (ICWITS), 2010 IEEE International
Conference.
Figure 7 Measured Reflection coefficient
202
Proceedings
Risk Area Mapping and Identification of Hotspots on

the Road-Network of Lahore
Spatial Pattern and comparision of Road Traffic Accident with population and commercial
area of district Lahore.
Uzair ul Hassan Shah
Muhammad Nauman Hussain
GIS Developer
The Urban Unit
Lahore, Pakistan
Uzairshah.gis@gmail.com
GIS Research Associate

The Urban Unit
Lahore, Pakistan
Nauman.gis11@gmail.com
Abstract if we talk about big cities first thing that comes to

our mind is crowded roads and accidents. We daily read it in
newspaper that many people died or injured were being rescued
by the Rescue teams but these type of news are becoming part of
our life, in other words its not strange for us anymore. Ratio of
accidents is increasing day by day because of many reasons like
over speeding, inexperience, intake of drugs, mobile phone use,
and negligence of drivers and to avoid the use of seat belts.
The current study is an effort of road accidents investigation
in Lahore district. For this purpose, road accident data of March,
2014, Lahore district is used that was stored by Rescue 1122. The
density function available in the spatial analysis tool of the Arc
Map 10.2 software was applied to identify the accident hotspots.
Both simple and kernel density tools were applied for the
assessment of risk areas on major roads of Lahore. Spatial
analyses for finding out risk areas of these road segments.
Furthermore a relation is drawn between the highly populated
area and the accident density. Also a comparison is made among
the road accidents and the commercials areas present in the
district. Based on the result, spatial distribution of Risks along the
Major road network of Lahore is marked on the map and
necessary measures are proposed to reduce the future accidents.
This geo-spatial data is vastly helpful for better planning of
combating strategies against road accidents.
Keywords Road Traffic Accidents, Road Risk Assessment,
GIS
I. INTRODUCTION
Transport system plays major role in the growth and
development of a nation to make safe transportation of persons
and goods from one location to another. Fatality/injury ratio in
road accidents in major cities of the world is alarmingly high.
Road accidents occasionally cause big tragedies by killing a lot
of people. There are many causes behind these accidents.
Persons travelling on the busy road are highly exposed to risk of
accident. Risk mapping has been carried out based on the
collected geo- tagged points and analyzed quality data of road
traffic accidents for managing further activities aimed at
improvement of traffic safety.
203
A significant precaution to save lives from the event of road

accident is provision of emergency/traffic control services at
appropriate distances. Limitations that explain the poor outcome
for people involved in road traffic accident in Pakistan have been
identified as inappropriate dedicated transportation and traffic
density is most in urbanized cities. Road traffic accident should
be examined also from spatial dimension by using GIS
techniques.
GIS (Geographical Information System) can be put in
effective use for accident analysis. GIS technology is used for
managing locational and relational information and it is able to
manipulate and visually display trends & data required for easy
comparison and decision making regarding future planning and
development of traffic system. Development of a GIS system to
analyze the traffic accident has been pursed towards improving
the effectiveness and efficiency of traffic accident.
II. STUDY AREA

Lahore is the capital of Punjab Province and is the most
urbanized and thickly populated district in Punjab province.
According to the 2013 census (Punjab Development Statistic
2013 Bureau of Statistic, Lahore) has 12,218,345 population
with 81.17% urban portion. The total area is 1,772 square
kilometer. Nearby Districts are Kasur, Nankana Sahib and
Sheikhpura. Major Roads in the District Boundary are National
Highway (N5), Ferozpur road, Grand Trank Road, Ring Road,
Canal Road, Mall Road, Jail Road, Multan Road, Allama Iqbal
Road, Raiwind Road, other primary and secondary roads with
different lengths,widths and Traffic Densities.
FIGURE I- THE URBAN AREA OF DISTRICT LAHORE WITH MAJOR ROADS
Proceedings
FIGURE II- MAP SHOWING THE ACCIDENT LOCATIONS IN DISTRICT LAHORE

DURING MARCH 2014
III. METHODS
D. Statisctis
A. Data Collection
In order to identify the accident locations in Lahore district,
following data were collected and used.
RTA (Road Traffic Accident) location data were collected
from the control room of Rescue 1122 (Emergency service)
in Lahore.
Accident Report for the month of march 2014
The analysis result is based on the available data collected
from Rescue 1122 databases of district Lahore. The Population
data (in the form of landscan image) is gathered from The Urban
Unit. That raster is further clipped according to the study area.
The commercial areas are identified by the Urban Unit. The
quality of data determines largely the quality of research result.
B. Digitizing
Digitizing is the process of encoding the geographical
features in digital form as X, Y coordinates on the georeferenced satellite image of district Lahore. Road network is
digitized as line feature and accident location as point feature in
Arc Map 10.1. The exact location of accident is identified by
using online Google maps, Bing maps and distance from the
relative location (Shops, Petrol pumps and traffic signals etc.)
by using distance measuring tool in Arc Map 10.1. The maps
showing the accident location in the figure II.
In the present study a general statistical analysis was

carried out in terms of accident occurrence on the primary and
secondary roads of Lahore at different times, locations and
causes of accidents. There are total 4528 accident cases in
rescue 1122database from March 01, 2014 to March 31,
2014.Among 4528 accidents, 718 pedestrian, 1315 passengers,
1998 Drivers and 325 under age drivers became victims
measuring 15.9%, 29.0%, 44.1% & 7.20% respectively of total
reported accidents.
TABLE 1: VICTIM DETAIL
Victim Detail
Accidents
Pedestrian
718
15.9
Passenger
1315
29.0
Driver
1998
44.1
Under Age Driver
325
7.20
In 4528 accidents: 2231 bikes, 365 cars, 14 busses, 39 trucks,

177 vans, 479 rickshaws and 277 other vehicles were involved
in the events, measuring 49.20%, 8.0%, 0.3%, 0.8%, 3.0%,
10.5% & 6.10% respectively of total reported accidents.
TABLE 2: VEHICLE DETAIL
C. Assigning Attributes
All vector data (i.e. line, point feature) had attributes. Here
roads were labeled with road name and total accidents on the
road segment that is from one intersection to another by spatial
join tool in Arc Map 10.1 point feature contain following
attributes:
Date of Accident occurrence.

Exact Time of occurrence.
Day of occurrence.
Place of accidents.
Vehicle Detail
Accidents
Bikes
2231
49.2
Cars
365
8.0
Busses
14
0.3
Trucks
39
0.8
Vans
177
3.0
Rickshaws
479
10.5
Others
277
6.1
Distribution of injuries as reported in the rescue database: 23

spinal injury cases, 335 head injury, 202 leg fractures, and 130
multiple fractures and 3341 minor injuries.
204
TABLE 3: DISTRIBUTION OF RTA INJURY
Proceedings
IV. RESULT AND ANALYSIS
Injury
Victim
Spinal Injury
23
0.5
Head Injury
335
7.3
Leg Fracture
202
4.4
Multiple Fracture
130
2.8
Minor Injury
3341
73.7
Compiling the data and performing statistical analysis shows

the high level of traffic accidents in this major city of Pakistan.
These statistical results are represented in the form of charts are
given below.
Causes of Road Traffic accidents reported in the database:

1756 over speed cases, 734 carelessness, 485 wrong turn, 328
U turn, 1 one wheeling, and 4 tyre brust, 274 others accidents
causes.
TABLE 4: DISTRIBUTION OF REASON CAUSE ACCIDENTS
Causes of RTA
Victim
Over speed
1756
38.7
Carelessness
734
16.21
Wrong Turn
485
10.7
U Turn
328
7.2
One Wheeling
0.2
Tyre brust
0.8
Others
1220
26.9
Patient condition on the spot: 16 patient died on spot 2761

victim were alive and unstable, 1756 alive and stable.
TABLE 5: DISTRIBUTION OF VICTIM CONDITION ON SPOT
Condition
Victim
Dead
16
0.35
Alive& unstable
2761
60.97
Alive & Stable
1751
38.67
Above charts show the RTA take place in March 2014. It

illustrates the randomness of the accidents in Lahore city. But
the charts did not explain the spatial feature of accidents i.e.
where was that accident happened. To show the location of
RTA, maps are produced which are displayed as follow.
Age of victims effected by the road traffic accident: 153 cases

of age 1 ~ 10Y, 854 were 11 ~ 20Y, 1389 were 21 ~ 30Y, 692
were 31 ~ 40Y, 485 were 41 ~ 50Y, 284 were 51 ~ 60Y, 174
were above 60Years.
TABLE 6: DISTRIBUTION OF VICTIM AGE
Age
Victim
1Y-10Y
153
3.79
11Y-20Y
854
21.18
21Y-30Y
1389
34.45
31Y-40Y
692
17.16
41Y-50Y
485
12.03
51Y-60Y
284
7.04
Above 60Y
174
4.31
FIGURE III-:RTA IN MARCH 2014
205
Proceedings
The above figure contain 4 maps demonstrate different time

zones of a day. A day is divided in to four time zones i.e. from
12AM to 6AM, 6AM to 12PM, 12PM to 6 PM, 6PM to 12AM.
Each map shows the density of accidents in the specific time
zone.
FIGURE IV-RTA DENSITY MAP
Kernel density is applied to get the strength of traffic accident

per unit area. In kernel density all those points are considered in
calculating density which are fall with in the neighborhood of a
particular cell. If no point fall within the neighborhood of a
particular cell, that cell is assigned no data.
FIGURE IVI- KERNEL DENSITY MAP
FIGURE VII- RELATION BETWEEN RTA AND POPULATION
The above map is the comparison of accidents and

population density. The population is represented in a tone of
red bright color which turns darker when we move towards the
denser population. The accidents density is also shown in atone
changing from green to red (red as most venerable area). The
comparison shows that most of the accidents occurs in highly
populated areas. Similarly accident strength is high on
major/main roads (in blue color lines) of the city.
FIGURE IV- RTA SPATIAL DISTRIBUTION W.R.T DAYS
The map shows spatial distribution of RTA w.r.t days of a

week. This shows the venerable areas of accident during
different days of a week which intends helpful for the traffic
authorities mitigate the venerability of accidents.
FIGURE VIII- RTA COMPARISON W.R.T COMMERTIAL AREA
FIGURE IVI- RTA DENSITY DURING DIFFERENT TIME ZONES OF A DAY
206
For Cities like Lahore having major trading points may have
risk of more RTA. To understand this problem above map
illustrate the RTA comparison with commercial area. The map
clearly shows that the majority of accident happened near
commercial areas. Another picture of same issue reflect the
comparison of RTA density with commercial points (Figure XI).
This map demonstrates that red zones for accidents are nearby
commercial areas.
Proceedings
traffic plans and rules. It also provides accurate

recommendations to vehicle drivers, the police and to local
monitoring authorities. In order to curtail road traffic accident
on Lahore roads the following recommendations are necessary:
Drivers and bikers should be trained as the mean of effective
dealing with road accidents.
Traffic rules must be deployed by the law enforcement
agencies in the district and country.
Road safety education and seminars should be held by the
transport departments.
Drivers should drive within speed limit and with a speed
unswerving with road condition.
There must have proper parking plazas in all major
commercial zones and markets.
VI. ACKNOWLEDGMENT
This study is sponsored by The Urban Unit (Urban Sector
Planning and Management Service Unit USPMSU). The authors
would like to thanks Mr Niazm-ud-din (GIS Manager at Urban
Unit) for providing data and Mr. Ehsan Saqib (Sr. GIS
Development Specialist), Dr. Ather Ashraf (Sr. GIS Specialist)
for helping in different analysis.
FIGURE IX- RTA DENSITY COMPARISON WRT COMMERCIAL AREA
V. CONCLUSION AND RECOMMENDATIONS
VII. REFERENCES
This work gives an insight of the present scenario of the

traffic condition in Lahore and shows out the most accident
prone roads. The result shows that total road traffic accidents,
dense population areas, commercial areas are the important
variables to take into the consideration in examining traffic
accidents. Spatial data can be shared with transportation,
emergency services and other government organizations. This
spatial database can be helpful for expert system to make new
[1]
[2]
[3]
[4]
[5]
207
Bureau of Statistic Bos, Punjab Development Statistics

2013,www.bos.gop.pk.
Deepthi Jayan.K, B. Ganeshkumar,Identification of Accident Hotspot: A
GIS Based Implementation for Kannur District, Kerala , vol. 1. ISSN
0976-4380.
Dr, A.J.Aderamo, Spatial Pattern of Road Traffic Accident Casualties in
Nigeria,, vol. III, ISSN 2039-2117.
Harnam Singh,A.D Aggarwal, Fatal Road Traffic Accident among
Young Children, J Indian Acad Forensic Med, 32(4), ISSN 0971-0973.
The Urban Unit, USPMSU Lahore, www.urbanunit.gov.pk
Proceedings
ANALYSIS OF RECENT DROUGHT BASED ON

NDVI AND METEOROLOGICAL DATA
A case study of Tharparkar district, Sindh
Muhammad Arslan1, Rao Zahid2, Badar Ghauri3
Department of Remote Sensing and GISc
Karachi, Pakistan
marslanhafeez100@gmail.com
Abstract Thar region of Pakistan has often undergone short

and long period of drought. In this study, Satellite Remote
Sensing and meteorological data were used to detect the drought.
Agricultural drought was identified based on the pre and post
monsoon Normalized Difference Vegetation Index (NDVI) with
250meter spatial resolution from MODIS terra satellite from
2008 to 2014. Departure from the mean values of pre and post
monsoon NDVI was categorized to determine the agricultural
drought. Whereas meteorological drought was identified based
on the annual precipitation data for the same period provided by
the Pakistan Meteorological Department (PMD). Pre and post
monsoon mean NDVI for the year 2014 was then compared with
the previous years. Pre monsoon mean NDVI values of 2008 to
2013 was 0.115956, while post monsoon mean NDVI values of
2008 to 2013 were around 0.221801. On the other hand, pre and
post mean NDVI for the year 2014 were observed as 0.127617
and 0.156314 respectively. Mean post-monsoon NDVI value
drastically dropped in the year 2014. Satellite based mean NDVI,
PMD annual rainfall data and field observation showed that a
short term drought occurred in year 2014. This was also
confirmed by the local population during the field visit.
Keywords Drought, MODIS, NDVI, precipitation.
I. INTRODUCTION
Drought is classified into four classes. Hydrological
drought, meteorological drought, agricultural drought and
socio-economic drought. No universal definition of drought is
still known. Drought is defined as a period of abnormally dry
weather which eventually affects in a change of vegetation of
an area [1]. Drought is also defined as it usually occurs when
the rainfall or runoff is found below a mean truncation level,
derived from the long term rainfall series [2]. For human
society and ecosystem drought is a significant adverse climatic
events. Since 1850, the mean global surface air temperature
has increased by 0.76oC [3], by the end of the twenty first
century, it is expected that the temperature will increase by 1.5
o
C 6.4 oC [4]. Drought may increase under the warm climate
[5]. The frequency and the intensity of drought have increased
in past decades in many regions of the world [6, 7]. This
recurring of drought has negative impact in these areas in term
of annual loss of vegetation and food. Drought monitoring has
gained importance in many countries especially developed

countries due to the serious social, environmental and
economic ramifications [8]. Conventionally, drought
monitoring was based on meteorological station observations
that obviously does not have a continuous spatial coverage for
the monitoring of comprehensive drought. Most of the studies
have used satellite data to monitor hazards including, drought,
flood, earth quake and so on [9, 10]. Satellite data have
frequently been used to monitor and detect the drought
phenomenon [1]. Satellite data provides earth observation for
monitoring the impacts of drought in a precise and effective
way. There exists a time lag between precipitation events and
its response on the vegetation. The time gap between the
occurrence of precipitation and the time precipitation water
reaches the plant roots. It varies from 01 to 12 weeks
depending on vegetation type [11]. There are various studies
that have focused on NDVI and precipitation relationship.
This relationship varies spatially, when the NDVI values
changes with respect to precipitation [12]. Spatial distribution
and natural condition of vegetation are the function of
environmental variables such as precipitation. Strong
relationship is occurred between NDVI and rainfall [11].
Precise and consistent vegetation mapping is essential and
necessary for the management of water resources and
mitigation of drought events.
A. Objectives
to identify the NDVI and meteorological based drought.
to compare 2014 year recent drought with the previous

years
to evaluate the effects of drought in term of socioeconomic condition of the local people
A. Study Area
Tharparkar district of Sindh province consists of deserted
land and its area is 19,638 square kilometers. Administratively
this district is divided into four talukas. These are Diplo,
Nagarparkar, Mithi and Chachro shown in Fig1. It has
208
population of 9, 14000 and density of 46.17 per square

kilometer.
III. RESULTS AND DISCUSSIONS
Figure 1. Map of the study area
B. Data Sources
Satellite data included Moderate Resolution Imaging
Spectro-radiometer (MODIS) 16 days composite of
Normalized Difference Vegetation Index (NDVI) from 2008
to 2014. Meteorological annual rainfall data of the duration of
Mithi observatory, Tharparkar was obtained from Pakistan
Meteorological Department (PMD). Side by side a field
survey was also undertaken in order to observe the drought
condition in this area.
TABLE 1. DATA USED AND ITS SOURCES

Data
Description
Sources
MODIS Product
NDVI
MODIS Satellite
Rainfall Data
Mithi Observatory
PMD
Study area Condition
Agriculture and Livestock
Field Observation
Proceedings
Pre monsoon mean NDVI value of 2008 to 2013 was

0.115956, while post monsoon mean NDVI of 2008 to 2013
was 0.221801. On the other hand, pre and post-monsoon mean
NDVI for the year 2014 was observed as 0.127617 and
0.156314 respectively. Post-monsoon mean NDVI value
dropped drastically in 2014 compared to post-monsoon mean
NDVI value for the period 2008-2013 shown in fig.3 and 4.
Annual rainfall (PMD data) also dropped drastically in 2012,
2013 and 2014 as 227mm, 189.1mm and 193 mm respectively
compared to 2011 which was 1310 mm (shown in fig.5). In
2014, annual rainfall increased a little by 4 mm, but still this
indicated drought in Tharparkar district. Lack of rainfall
caused loss of natural vegetation. Quantitatively, livestock
directly depends on the natural vegetation for fodder (as
shown in fig.6). In Tharparkar, agriculture activity mainly
depends on the mercy of rainfall. During the field survey
carried out, no agriculture activity was observed because
drought was the main reason affecting local agriculture
(shown in fig.7). Locals were waiting for the rainfall and they
left their ploughed land in expectation of rain. The satellite
NDVI, PMD annual rainfall and field based observations
showed that a short term drought might have occurred due to
the lack of rainfall compared to 2011 and a loss of natural
vegetation in post monsoon season of 2014. The 2014 drought
was also confirmed by the local population during field visits
shown in Fig 8(d). Most of the vegetation cover reduced and
thus developing a seasonal drought 2014 in Tharparkar. Based
on satellite, field and weather data, 2014 drought had occurred
in Tharparkar District, Sindh.
C. Methodology
Following were the main steps taken to accomplish the
research study and as illustrated in Fig.2.
After having acquired all the necessary data (described
above), pre-processing of the data was done.
Mean NDVI of the pre and post monsoon seasons were

used for the study.
Information on fodder and trend of agriculture activities
collected from the field of Tharparkar district on February
2015.
MODIS mean NDVI, field observations and PMD annual

rainfall data were integrated in order to analyze the drought
severity condition specifically for the year 2014.
Figure 3. Post monsoon mean NDVI (2012-2014)
2008
2009
2010
2011
2012
Figure 4. Pre and post monsoon mean

NDVI values (2008-2014)
Figure 2. Methodological framework chart
209
2013
2014
Proceedings
IV. CONCLUSION
Drought is a recurring phenomenon and it can be examined
by satellites. Long term satellite data could help the severity
and onset of drought due to its spatial and temporal variability.
Effective and efficient drought mitigation could be achieved
through a well-defined drought monitoring system.
Figure 5. Annual rainfall provided by PMD (2009-2014)
ACKNOWLEDGMENT
This research is part of a project named Development of
Decision Support System for Drought Monitoring in Sindh
funded by the Pakistan Higher Education Commission (HEC).
The authors acknowledged the financial grant of HEC.
Thanks are also due to Pakistan Meteorological Department
(PMD) for provision of data. We also thank NASA site for
giving data.
REFERENCES
Figure 6. Means of food for the livestock
[1]
Kogan, Felix N. "Global drought watch from space." Bulletin of the

American Meteorological Society 78.4 (1997): 621-636.
[2] Dracup, John A., Kil Seong Lee, and Edwin G. Paulson. "On the
statistical characteristics of drought events." Water resources research
16.2 (1980): 289-296.
[3] Jones, P. D., Trenberth, K., Ambenje, P., Bojariu, R., Easterling, D.,
Klein, T., ... & Zhai, P. (2007). Observations: surface and atmospheric
climate change. In Climate Change 2007: The Physical Science Basis.
Contribution of Working Group I to the Fourth Assessment Report of
the Intergovernmental Panel on Climate Change, 235-336.
[4] Solomon, Susan, ed. Climate change 2007-the physical science basis:
Working group I contribution to the fourth assessment report of the
IPCC. Vol. 4. Cambridge University Press, 2007.
[5] Pachauri, Rajendra K., and A. Reisinger. "Synthesis Report." IPCC.
Ginebra (2007).
[6] Hulme, Mike, and Mick Kelly. "Exploring the links between
desertification and climate change." Environment: Science and Policy
for Sustainable Development 35.6 (1993): 4-45.
[7] McCarthy, James J. Climate change 2001: impacts, adaptation, and
vulnerability: contribution of Working Group II to the third assessment
report of the Intergovernmental Panel on Climate Change. Cambridge
University Press, 2001.
[8] M Amin, Owrangi, et al. "Drought monitoring methodology based on
AVHRR images and SPOT vegetation maps." Journal of water resource
and protection 2011 (2011).
[9] Anderson, James Richard. A land use and land cover classification
system for use with remote sensor data. Vol. 964. US Government
Printing Office, 1976.
[10] Peters, Albert J., et al. "Drought monitoring with NDVI-based
standardized vegetation index." Photogrammetric engineering and
remote sensing 68.1 (2002): 71-75.
[11] Chopra, Parual. "Drought risk assessment using remote sensing and GIS:
a case study of Gujarat." ME thesis, Indian Institute of Remote Sensing,
Dehradun & International Institute for Geo-information & Earth
Observation, the Netherlands (2006).
[12] Nicholson, S. E., and T. J. Farrar. "The influence of soil type on the
relationships between NDVI, rainfall, and soil moisture in semiarid
Botswana. I. NDVI response to rainfall." Remote Sensing of
Environment 50.2 (1994): 107-120.
Figure 7. Reduction in agriculture
Figure 8. (a and b) Livestock of the area, (c) How local people collect water
(d) Interview from the local people related to drought (e and f) Showing
local people in the picture
210
Proceedings
Assessment of Urban Growth of Karachi: From A

Tiny Town to A Meta City of the World
Ali Jan Hassan1, Mudassar H. Arsalan2, Hira Fatima
Institute of Space and Planetary Astrophysics
University of Karachi
Karachi, Pakistan
1
Alijan.shaaan@gmail.com
2
mharsalan@uok.edu.pk
Abstract Karachi is the only meta city of Pakistan and capital

of Sindh. It plays a key role in the countrys economy by
contributing 65% of total revenue. However the sustainability of
this city of lights is at risk due to mismanagement and
development without proper long term planning. The increased
urbanization has brought several severe issues such as increased
demographic density, management challenges, resource deficit
and ecological pressures, etc. The sprawl of the city is
uncontrolled spreading of urban developments. Remote sensing
data and GIS are now widely being used for change detection
analysis and applications. Urban sprawl is one of them, which
can be analyzed efficiently and cost effectively by utilizing
satellite data available online in the form of present as well as
historical images in different spatial resolutions.
In this study change detection technique was used to assess the
urban sprawl through the integration of historical maps with
Landsat-5 TM sensor image of 1989 and Landsat-8 OLI images
of 2014. With the aid of these images we monitored, measured
and analyzed the urban sprawl of Karachi. Despite of many
unsustainable conditions in the city, its population growth is
tremendous i.e. approximately 4 % per annum. This growth is
the result of the high natural increase as well as substantial
migrations from other parts of the country. It is assessed that the
living population in the mid 2014 was around 21 million, which
was far greater than of 1998 i.e. 9.2 million. In the last 25 years,
Karachi has extended around 15 percent at the rate of 2 square
kilometers per year to accommodate its increasing population
with considerable high density.
Index Terms Urban sprawl, Change detection, Sustainability,
Satellite imagery, GIS, RS.
I. INTRODUCTION
Sustainability is not a new concept, it was discussed,
analyzed and even practiced at all levels, from country to
individual. The term Sustainable Urban Developed (SUD)
emerged from the Brundtland Report [1], according to this
report, SUD is defined as Development that meets the needs
of today without compromising the needs of the future. Many
organizations such as UN-Habitat, WHO, EEA, etc., started to
work on SUD indicators. Indicators for sustainable
development provide valuable information, indicating rather a

development is on the way of sustainability or degradation [2].
Planners, developers and stakeholders use the SUD indicators
to better understand the current development situation as well
as to predict the future development scenario in context of
sustainability[3] Urban Sprawl is one of the main indicators of
SUD, that can determine the increment area of Urban land, as
well as the densification of the city [4]
Urbanization and Urban Sprawl are two different
terminologies. Urbanization is a process of increasing number
of people that migrate from rural to urban areas, results in the
physical growth of urban areas regardless of developments
occur horizontally or vertically [5], whereas Urban Sprawl is
defined as the expansion of urban area in the form of new
development on isolated tracts, separated from other areas by
vacant land [6]. The socioeconomic conditions of any place are
the driving forces for urbanization hence it occurs more
tremendously in areas that are economically well established
[7]. Urbanization is an inevitable process due to economic
development and rapid population growth [8]
In the last part of the 20th century and the beginning of the
21st century human population has increased tremendously. On
average, at every second 4 or 5 children are born, while 1 or 2
people dies, this rate of birth and death means a net gain of 2.5
persons (on average) on the Earth [9]. Unfortunately, this
increase has been occurring in the developing part of the world,
where lack of facilities, poverty, pollution and crime rate are at
its peak [9].
Pakistan is in the list of developing countries and is badly
affected by terrorism, energy crises and law and order situation
in the country [11]. It is an atomic power, bearing worlds
toughest army, having a vast amount of resources of coal, oil
and other minerals [12]. But due to miss-management,
irregularities
and
irresponsibleness of bureaucracy,
stakeholders and political leaders, the complete potential of the
resources is not being properly utilized [13]. Pakistan is the
worlds sixth most populous country today [14] With an
estimated population of 188 million, and 2.05 percent annual
211
growth rate, Pakistan is expected to become the fifth largely

populated nation by 2050 [14].

The simple methodology adopted for this study comprises
of three steps: Data collection, data processing and urban
sprawl analysis. The brief account of these steps is given
below.
Table 1. Worlds Most Populous Cities [14]

2014
2015
Country
population
(Millions)
Country
population
(Millions)
China
India
USA
Indonesia
Brazil
Pakistan
1368
1226
320
255
203
188
India
China
USA
Nigeria
Pakistan
Indonesia
1628
1437
420
299
295
285
Proceedings
A. Data Collection
This study is an application of remote sensing and GIS in
Urban sprawl analysis. As recent as well as historical remote
sensing data is freely available on the internet. Further, using
GIS on this Remote sensing data different analysis can be
done efficiently and cost effectively. For this study Landsat 5
TM sensor clod free image of Karachi (Path-Row 152-42)
captured in November 1989 and Landsat 8 OLI sensor image
captured in January 2014 were acquired from the Landsat
archives available at the USGS website.
B. Data Processing
The pre-processing was carried out for the development of
AOI. Downloaded images of the study area were
geometrically corrected and projected to Universal Transverse
Mercator (UTM) projection system, zone 32. The spheroid
and datum were referenced to WGS1984.
Fig.1. Population growth in developing and developed nations, 1750

2100 [10]
A. Study Area
Today Karachi is the largest and most populous city of
Pakistan, lies in the south bounded by the Arabian Sea. It
covers a total area of 3600 square kilometers that includes
Built-up land, Vegetation Coverage, Industrial built-up,
commercial areas and a lot of recreational areas. The Karachi
port serves as a gateway for all the import/export activity by
sea. The industrial and economic activity in Karachi support
Pakistan a 65% of revenue [15]. Karachi appeared as a megacity with the beginning of this century with an exponential
growth in its population.
The residents of Karachi are blessed with the sea breeze
that blows throughout the year. Because it is close to the Sea
the temperature remains normal, the mean annual temperature
is in between, the wind normally blows towards North-East.
There are many commercial centers exist in the city, which
enabled the it to develop as an important industrial and
economic center that attract the people not only from Pakistan
but all over the world.
B. Objectives
The Objective of this study was to determine the urban
footprint for the year 1989 and 2014 using remote sensing data
and GIS, to figure out the extension of Built-up land within 25
years (1989-2010).
Fig. 2. Study Area- Karachi, Pakistan
C. Methodological Framework for Urban Sprawl

The change detection technique was adopted to assess the
change in land use and land cover (LULC) of the study area.
Collected and processed Landsat 5 image of 1989 and Landsat
8 image of 2014 were used while the main focus was on built-
212
up area change. To extract the Urban built up the more

accurately supervised maximum likelihood classification was
done.
The Maximum Likelihood Classification is one of the
supervised classification approaches for pattern recognition in
which the probability of a pixel belonging to each of a
predefined set of classes is calculated, then pixel is assigned to
the class for which the probability is the highest [16]. ML is
based on the Bayesian probability formula:
P(x, w) = P(w|x) P(x) = P(x|w) P(w)
Where x and w are called events, P(x, w) is the probability
of coexistence of events x and w, while P(x) and P(w) are the
prior probabilities of events x and w and P(w|x) is the
conditional probability of event x given event w [16].
By carefully observing the patterns in the images, it was
concluded that the classification should be done for 4 classes
i.e. Urban Land, (including dense, green as well as sparse
Urban areas), Vegetation (including dense and sparse
vegetation as the study area had a vast mangrove forest towards
its south), Water bodies and Open/barren land.
Maximum Likelihood Classification was done on natural
color composite image (Band Combination 3, 2, 1) by training
100 samples of four classes distributed randomly in the image.
Classes were identified by visual interpretation. From this a
thematic map of four classes was generated. The same
procedure was applied to the other image.
Proceedings
was also used to assess the difference between actual

agreement and change agreement [16].
III. RESULTS
Due to the highly favorable geographical location, Karachi
has always been a trade route for the merchants of India, but it
gained its importance in the British rule. The development of
perennial irrigation schemes by the British in Punjab and Sindh
mainly accounted for the growth of Karachi along with that the
development of railways, making the quick transportation of
the agricultural material that in turn increased in agricultural
production which was exported through Karachi which made it
the largest exporter of wheat and cotton to India [17].
In 1935, when Sindh separated from Bombay, Karachi
became the capital of Sindh, Government offices and trade
organizations were shifted from Bombay to Karachi. In 1947,
Pakistan got independence and Karachi became first capital
Pakistan.
Table 2 Karachi Population Growth [18]
Year
Population
Relative
change
1901
1911
1921
1931
1941
1951
1961
1972
1981
1998
2001
2005
2014*
136,297
186,771
244,162
300,779
435,887
1,137,667
2,044,044
3,606,746
5,437,984
9,802,134
13,500,000
15,119,000
20,228,112
136,297
323,068
430,933
544,941
135,108
701,780
906,377
1,562,702
1,831,238
4,540,422
3,697,866
1,619,000
5,109,112
No. of year in
between two
consecutive
census
10
10
10
10
10
10
10
11
9
17
3
4
9
Population
change
(%)
37
30.2
23.2
44.9
161
79.67
76.5
50.8
86.29
37.7
11.9
25.5
From the below graph we can see that the population of

Karachi was 9.2 million in 1998, there is a tremendous increase
of around 12 million in 2014[18].
Fig. 3. Methodological Framework of Urban Sprawl
1) Accuracy Assessment
To check the level of confidence in the classification results
an error matrix is usually generated. This is some time called
the confusion matrix, in which a matrix presents the
relationship between the classes in the reference maps and
classified images. Estimating classification accuracy enables us
to determine classification performance [16], but there is a lack
of information, because it leads to overall accuracy of the
classes not of individual class. To assess the accuracy of
individual class the concept of Producers and Users accuracy
was used in this paper. A multivariate index, Kappa Coefficient
213
Fig. 4. The Population Trend of Karachi 1901-2014
Proceedings
There was a significant increase of 161% from 1947-1951,

because of the migration of 600,000 refugees from India. The
settlement of those refugees was a challenge for newly built
local government. The refugees started to live on the periphery
and within the city itself, as a result the environment of Karachi
totally changed not only demographically but also ethnically
[17]. The irresponsibilities of higher authorities and lack of
planning and strategy the refugees unwillingly started to live in
squatter settlements and katchi abadis. After 1958, five master
plans have been developed for the development of the city, but
the implementation of those plans has always been an issue
because of political instability within the country.
The Karachi City, administratively is spread over an area of
approximately 3,600 sq. km. More than half of this area
(approximately 2455.5 acres) consists of vacant land. The landcover features such as water and vegetation, changes
throughout the year because it depends on the Climatic
conditions at a particular time. The Urban Built-up area has
extended from 343 square km in 1989 to 396 square km in
2014 that is a valuable change of 53 km in 25 years. This
reason behind this rapid increase in population within the city
is the migration from Interior Sindh, Punjab and Khyber
Pakhtoonkhuwa to find better opportunities in the mega city.
There is a 100 square km decrease in open land from which 50
sq. km land is converted into built-up land
The actual built-up land is 396 square km but as per
document figures by the City District Government of Karachi
(CDGK) [19], the built-up land was 1300 square km, it was a
combined built-up and developed open spaces. Developed
open spaces are those housing societies which were initiated in
1970s, 80s, and 90s, however, even after their infrastructure
developed they are vacant and hardly 5% occupied as shown in
table 3.
Fig. 9. Land Cover change % of Karachi (1989 - 2014)
Table 3 Occupancy Status of New Housing Scheme, [19]

Serial. No.
Name of
Scheme
Year of
Notification
Scheme no. 25A

Scheme no. 33
Scheme no. 42
Scheme 43
Scheme 45
1980
Current
Occupancy
Status
(%)
5%
1971
1983
1986
1986
20%
5%
0%
5%
2
3
4
5
Fig. 8. Land Cover Change of Karachi (1989-2014)
Table 5 Error Matrix of Land Cover (Karachi - 1989)
From the table below, we can see that there is a 15 %

increase in built-up land and 3% decrease in open land. This
growth is mainly in built-up area and decrease in open land.
Table 4 Area Cover by each class for the years 1989 and 2014 in Sq. km
Serial
No.
Class
1
2
3
4
Built-up
Open Land
Vegetation
Water
Area
(1989)
Sq. km
343
2556
748
1014
Area
(2014)
Sq. km
396
2455
849
972
Change Change
Sq. km
(%)
53
-100
-100
-41
15
-3
13
-4
Class
Builtup
Open
Land
Veget
ation
Wat
er
Row
Sum
User's
Accura
cy
Builtup
Open
Land
Veget
ation
Water
Colu
mn
Sum
14
15
93.33
Prod
ucer's
Accu
racy
93.33
45
55
81.82
100
10
10
100
45.45
0
15
0
45
2
22
18
18
20
100
90
100
Overall Accuracy 87%
214
Kappa 0.328
Proceedings
Fig. 10. Growth of Karachi 1938-1992 [20]
Table 6 Error Matrix of Land Cover (Karachi - 1989)

Class
Builtup
Open
Land
Veget
ation
Wate
r
Row
Sum
User's
Accurac
y
Produ
cer's
Accur
acy
Builtup
Open
Land
Vegeta
tion
Water
Colum
n Sum
12
66.67
88.89
44
52
84.62
88
100
64.29
0
9
2
50
0
14
25
27
27
100
92.59
92.59
Overall Accuracy 86%
Kappa 0.3563
IV. DISCUSSION AND CONCLUSION

This study reveals that according to CDGK (2007) the
population of the city reaches to around 21 million and if we
consider the whole of Karachi area that is 3,600 km, the
population density is 6000 persons per square kilometer about
25% of its population has increased in last 10 years. However,

the Urban built-up area has extended from 343 square km in
1989 to 396 square km in 2014 that is a change of 53 square
km in 25 years. It shows the average annual growth rate is
slightly above 2 square kilometers per year. If we consider
average population density is 50,000 per square kilometer, the
city can accommodate only 0.1 million people every year.
However, it is not so, as the average population increase is 5
times higher than its accommodating capacity. It indicates the
pressure of population is more towards the existing
neighborhood in the form of densification and slum area's
growth. That results in depletion of resources, degradation in
environmental conditions and lack of recreational areas within
the city center. All these factors produce bad impacts on
social, environmental and economic conditions of the city and
makes difficult to achieve sustainable development.
This study was conducted to assess the conditions of Urban
Sprawl in Karachi that is an effective indicator of SUD,
however many other indicators can also be identified and
assessed, by implementing the techniques of RS and GIS.
215
Fig. 10. Land Cover Map of Karachi 1989
Fig. 11. Land Cover Map of 2014
216
Proceedings
REFERENCES
[1] World Commission, "Report of the World Commission on
Environment and Development: Our Common Future Acronyms
and Note on Terminology Chairmans Foreword," ed: Oxford
University Press, 1987.
[2] M. Javadian, H. Shamskooshki, and M. Momeni, "Application
of sustainable urban development in environmental suitability
analysis of educational land use by using AHP and GIS in
Tehran," Procedia Engineering, vol. 21, pp. 72-80, 2011.
[3] N. Chrysoulakis, C. Feigenwinter, D. Triantakonstantis, I.
Penyevskiy, A. Tal, E. Parlow, et al., "A Conceptual List of
Indicators for Urban Planning and Management Based on Earth
Observation," ISPRS International Journal of Geo-Information,
vol. 3, pp. 980-1002, 2014.
[4] Sustainable Cities International, Indicators for sustainabilityHow cities are monitoring Sustainability, 2012.
[5] UN, "United Nations Conference on Environment &
Development, Rio de Janerio, Brazil, 3 to 14 June 1992," 1992.
[6] J. I. IGBOKWE, I. C. EZEOMEDO, and E. J, "Investigation of
Urban Sprawl Using Remote Sensing and GIS: A case of
Onitsha and its Environments in Southeast, Nigeria,"
International Journal of Remote Sensing & Geoscience (IJRSG),
vol. 2, 2013.
[7] Boundless. (2015, July). The Process of Urbanization.
Available:
https://www.boundless.com/sociology/textbooks/boundlesssociology-textbook/population-and-urbanization17/urbanization-and-the-development-of-cities-123/the-processof-urbanization-695-3433/
Proceedings
[8] S. Deep and A. Saklani, "Urban sprawl modeling using cellular

automata," The Egyptian Journal of Remote Sensing and Space
Science, vol. 17, pp. 179-187, 2014.
[9] W. P. Cunningham and M. A. Cunningham, Principles of
environmental science: inquiry & applications: McGraw-Hill,
2011.
[10] D. B. Botkin, E. A. Keller, and D. B. Rosenthal, Environmental
science: Wiley, 2012.
[11] I. Rana, "Foreigners driven away by terrorism, energy crisis," in
The Express Tribune, ed: @etribune, 2011.
[12] H. Naz, Environment of Pakistan, 2006.
[13] I. A. Sodhar. (2011, July). Pakistan Rich in Natural Resources
But
Poor
in
their
Management.
Available:
http://jworldtimes.com/Article/92011_Pakistan_Rich_in_Natura
l_Resources_But_Poor_in_their_Management
[14] U. S. C. Bureau. (2011, July). International Programs, Country
Rank.
Available:
https://www.census.gov/population/international/data/countryra
nk/rank.php
[15] A. Hassan. (2015, July). Ishaq Dars budget defies economic
vision: Karachi Chamber of Commerce and Industry. Available:
http://asianetpakistan.com/business-news/general-businessnews/203382/ishaq-dars-budget-defies-economic-visionkarachi-chamber-of-commerce-and-industry-3/
[16] P. Mather and B. Tso, Classification methods for remotely
sensed data: CRC press, 2009.
[17] A. Hasan and M. Mohib, "The case of Karachi, Pakistan," 2003.
[18] Government of Pakistan, "Population Census Reports."
[19] CDGK, "Karachi Strategic Development Plan 2020," 2007
[20] M. S. Faruqui, Karachi, Physical Situation of Human
Settlements: Master Plan and Environmental Control
Department, Karachi Development Authority, 1982.
217
Proceedings
JUPITER, THE GAS GIANT

ABDUR RAQEEB GAZIANI, POSTER COMPETITION
Jupiter is the first of gas planets in our solar system. It travel around sun once every
11.86years. Jupiter is so large that over 1300 earths can be packed in its volume. Here we will
discuss some bold points of Jupiter:
THE SURFACE:
Is mostly of gaseous hydrogen and the layer is about 20,000 km thick. Jupiter is one
of the brightest plant in the sky with the great red spot on surface. Convection currents
and the Jupiters rapid rotation create the bright zones and dark belts.
MOONS OF JUPITER
There are at least 63 natural satellites or moons orbit Jupiter. The four major moons of
Jupiter:
o Io: Io is about the same size as Earths moon. It orbits Jupiter once every 1.77
days. Io has nine giant erupting volcanoes on its surface and up to 200 smaller
volcanoes. There are also mountains up to 10 km high but these are not
volcanic.Io is thought to be volcanically active because of huge high tidal
forces created by Jupiter.
o Europa: Europa is the fourth largest moon of Jupiter. It orbits Jupiter once
3.55 days Europa has a thin outer layer of ice and a layered internal structure,
probably with a metallic core. Its surface is relatively smooth, with no
mountains and very few craters. Some astronomers think that a layer of liquid
water may be present below the ice-covered surface.
o Ganymede: Ganymede is the largest moon in solar system. Its diameter is
largest than the planet mercury. It orbits Jupiter in synchronous rotation once
every 7.16 days at a distance of about one million km.The crust is thought to
be 75 km thick and contain an outer layer of ice. The Hubble space telescope
has recently found evidence oxygen on Ganymede.
o Callisto: It is second largest moon orbiting Jupiter. It is just slightly smaller

than planet Mercury, but has only one-third its mass.Callisto orbits Jupiter
once every 16.69 days. Its surface is a dark, ancients and icy crust, covered
with many old impact craters. It probably cooled very rapidly. Voyagers
instrument measured a temperature range from -180C during day time to 315C at night. It has little internal structure, composed of about 40% ice and
60% rocky iron.
MAGNETIC FIELD:
Jupiters magnetic field is about 20 times stronger than Earths. Its magnetosphere
extends only a few million kilometres towards sun, but it extend more 650 million
kilometres away from sun. Jupiters magnetosphere is pushed by the solar winds. The
magnetic poles are offset from the rotational axis by nearly 10.
DENSITY AND COMPOSITION:
Jupiter is more than 318 times more massive than the Earth and has twice as much
mass as all the other planets combined. Jupiters average density is lower than that of
Earth only 1.33g/cm compared to Earths 5.52g/cm.
218
Proceedings
Identification of Post Disaster Scenario Using

Double Threshold Energy Detection
Shakeel Ahmed Waqas*, Fahad Shamshad
Nisar ali
Electrical Engineering Department, Military College

of Signals, National University of Science and Technology,
Islamabad, Pakistan
*
Shakeel_aw@yahoo.com
Department of Telecommunication Engineering,

Peshawar, Pakistan
Wireless Emergency Network making it more stable, energy

efficient and reliable. In [3], authors presented cross layer
framework for the post disaster scenario. The authors
presented an adaptive protocol selector which select
appropriate protocol for each layer. Furthermore, the authors
in [4] proposed three algorithms using single threshold based
on cognitive radio, cellular network and artificial neural
network to detect the post disaster scenario. But there is
possibility of false alarm if the mobile is in deep fading area.
The authors in [5], proposed double threshold energy detection
algorithm for cognitive network and proved that false alarm
can be reduced by using double threshold.
AbstractThe disaster can be naturally or manmade.

Telecommunication cellular system can be completely
destroyed by this disaster. The main priority is to provide
connection to disaster areas before the rescue team are
being arrived and save as much lives as possible. In this
paper, we presented a new algorithm which enable mobile
station to automatic detect the disaster scenario and switch
to emergency mode. We used double threshold energy
detection to reduce the false alarm. The performance of
proposed algorithm is evaluated using Monte Carlo
Simulation in term of probability of detecting post disaster
scenario, probability of miss detection and probability of
false alarm.
Keywords Post Disaster Communication, Emergency
Mode, Normal Mode, Double Threshold, Energy Detection

I.
INTRODUCTION
The telecommunication existing network can be

damaged or destroyed by manmade or natural disaster. In
result, that area will have no connection with national
communication infrastructure. 50% death occurs in two hours
after disaster [1]. The lives can be saved if there are some
mean of communication before the arrival of rescue team. The
survivors need urgent help after the disaster. Therefore, post
disaster communication is required to connect the disaster area
with whole country [1].
The post disaster communication can be achieved by
using Ad-hoc network. Ad-hoc network is used by many
researchers for emergency network. Wireless Emergency
Network can be obtained by multi hop Ad-hoc network. In
Wireless Emergency Network, there is ability of self-healing
[2]. But in disaster scenario, the resources are limited. Ad-hoc
network cannot reuse the available resources and also
unstable. In [1], the authors proposed that using cluster based
Ad-hoc network, the network become stable and resources can
be reuse. Thus, it increase the capacity. In cluster based
network, cluster head will forward the data of cluster
members. In [2], the authors proposed route repairing for
219
In the paper, we presented a new algorithm using

double threshold for detecting post disaster scenario. We used
energy detection technique for the decision making. In this
paper, we presented the performance of our proposed
algorithm using BPSK modulation scheme over AWGN
channel. The performance is analyzed in term of probability of
detection post disaster scenario, probability of misdetection of
post disaster scenario and probability of false alarm. All the
results are evaluated using Monte Carlo simulation in
MATLAB.
II. SYSTEM MODEL
The system investigated in this paper is shown in
figure 1. Our system model consists of one mobile station and
two base stations. The mobile station is denoted by MS and
the two base stations are denoted by BS1 and BS2. MS is
connected with BS1 and BS2 is adjacent base station. The
distance between the MS and BS1 is denoted by d1. Similarly,
the distance between the MS and BS2 is denoted by d2. In our
system model, d1 is 2 km and d2 is 3 km. AWGN channel is
considered between MS, BS1 and BS2. The path loss
exponent is considered as n = 2. All the nodes (MS, BS1 and
BS2) are mounted with single antenna.
It is assumed that MS has already installed software
based identification module. There are two modes on which
MS can operate. One is Emergency Mode and second is

Normal Mode. In Normal Mode, the MS will operate normally
d1
If Vconnected is less than
0 and Vadjacent is greater
than 1 , again it will be uncertain
BS 2
BS1
AWGN
5.
Proceedings
AWGN
d2
Normal Mode
Uncertain
Emergency Mode
Energy
Fig2. Double threshold Energy Detection algorithm for post disaster

identification
The advantage of our algorithm is that it reduce the

false alarm. If the MS is in deep fading areas where the signals
are weakly received, the MS will not switch to Emergency
mode. According to our proposed algorithm, the MS cannot
change its mode easily. It will change its mode when the
energy of connected BS and adjacent BS both are higher than
threshold 1 or lower than threshold 0 .
MS
Fig1. System Model
and will be connected with its mobile service provider

infrastructure recognizing itself as a part of cellular network.
In Emergency Mode, the MS will stop broadcasting messages
to BS and switch to Ad-hoc mode.
IV. RESULTS AND DISCUSSION
Switching the MS from one mode to another, MS

regularly sense the signal coming from connected base station
BS1 and adjacent base station BS2. Here we used energy
detection technique and double threshold to take decision and
switch the MS to appropriate mode. The detected energy from
BS1 is denoted by Vconnected and the detected energy from BS2
is denoted by Vadjacent .
The system model shown in previous section is simulated

using Monte Carlo simulation. Results are discussed in this
section. The time bandwidth factor is taken as 2. Average SNR
(signal to noise ratio) is 14 dB. The path loss is considered. The
path loss exponent is 2. Modulation used for transmission is
BPSK. The distance between MS and BS1 is 2 km and distance
between MS and BS2 is 3 km. AWGN noise is considered in
channel.
III. DETECTION ALGORITHM

Our proposed algorithm uses two thresholds i.e.
0 and 1 . The decision making is totally depends on detected
energies Vconnected & Vadjacent and values of threshold.
Algorithm:
1. If the Vconnected and Vadjacent both are greater than 1
, then switch to Normal Mode.
2.
If the Vconnected and Vadjacent both are less than

then switch to Emergency Mode.
3.
If the Vconnected and
0 ,
Vadjacent both are between
0 and 1 , then dont take any decision. Scan again

4. If Vconnected is greater than 1 and Vadjacent is less
than 0 , in this case, no decision would be taken.
Scan again to take better decision
Fig3. Probability of detection post disaster of scenario
The figure3 show the probability of detecting post disaster

scenario for different threshold 0 . As clear from simulation
result that for 0 dB 0 , the probability of detecting post
disaster scenario is zero. Means we MS will not switch to
Emergency mode. As we increase the value of 0 , the
220
probability increases. The post disaster scenario can be easily

detected if 0 is high.
Proceedings
V. CONCLUSION
In this paper, we presented double threshold energy
detection algorithm for detecting post disaster scenario. This
algorithm enable mobile station to detect the post disaster
efficiently and take right decision in switching its mode from
normal to emergency or emergency to normal. Decision is
made by considering the energy of signals received from
connected Base station and from adjacent base station. The
performance is presented in term of probability of detection of
post disaster scenario, probability of miss detection of post
disaster scenario and probability of false alarm. In future, we
will work future on identification of post disaster scenario
using double threshold considering the adjacent channels.
The probability of miss detection of post disaster scenario

is shown in figure4. From the result, it is obvious that we will
surely miss the detection of post disaster scenario if the
threshold 1 is kept very low. Keeping the 1 low, then greater
probability of switching to Normal Mode and low probability
of switching to Emergency Mode. As the threshold 1
increases, then we can easily detect the post disaster scenario.
Therefore, the probability of miss detection decreases as the
threshold 1 increases.
The probability of false alarm of our proposed algorithm is
shown in figure5. To calculate false alarm we assume that SNR
REFERENCES
[1]
[2]
[3]
[4]
Fig4. Probability of miss detection of post disaster

scenario
[5]
[6]
[7]
[8]
Fig5. Probability of false alarm

is 0 dB. According to our simulation result, it is stated that if
the thresholds are low then the probability of false alarm
increases. We get 100 % probability of false alarm at 0 dB
threshold.
As the threshold increases, the probability of false alarm
decrease. We get 0% false alarm for threshold equal to or
greater than 10 dB.
[9]
221
Sonia Majid and Kazi Ahmed, Cluster-based

Communications System for Immediate Post-disaster
Scenario Journal of Communication, Vol. 4, No. 5, June,
2009.
Xin Ma, Quan-yi Huang and Yan Kang, A New Algorithm
for Stable Communication in Wireless Emergency
Network International Conference on Computer and
Management, pp. 1-4, May 2011, Wuhan, China.
Sonia Majid and Kazi Ahmed, Cross-layer Framework
for Post-disaster Communications IEEE International
Conference on Telecommunications and Malaysia
International Conference on Communications, 14-17 May
2007, Penang, Malaysia
Sonia Majid and Kazi Ahmed, CIP- Cognitive
identification of post-disaster communication IEEE
conference on Telecommunications, pp. 43-47, June 2012,
Thailand
Jinbo Wu, Tao Luo, Jianfeng Li and Guangxin Yue A
Cooperative Double-threshold Energy Detection Algorithm
in Cognitive Radio Systems IEEE conference on Wireless
Communications, Networking and Mobile Computing,
pp.1-4, Beijing, China.
Haitao Wang, Lihua Song, Jianzhou Li, Jiaxin Deng,
Wireless Self-organizing Emergency Communication
Network based on Network Clustering and Information
Ferry , 2nd Internation Conference on consumer
Electronics, Communications and Networks, pp. 37043707, Yiching, China.
Sonia Majid and Kazi Ahmed, Post-disaster
Communications: A Cognitive Agent Approach 7th
International Conference on Networking, pp- 23-30 Cancun
Mexico, April 2008.
Marco Di Felice, Luca Bedogni, Angelo Trotta, Luciano
Bononi, Fabio Panzieri, Giuseppe Ruggeri, Gianluca Aloi,
Valeria Loscr, Pasquale Pace, Smartphones Like Stem
Cells: Cooperation and Evolution for Emergency
Communication in Post-Disaster Scenarios First IEEE
International Black Sea Conference on Communications
and Networking 2013 (IEEE BlackSeaCom 2013), Batumi,
Georgia.
Jinbo Wu, Tao Luo and Guangxin Yue, An Energy
Detection Algorithm Based on Double-threshold in
Cognitive Radio Systems The 1st International Conference
on Information Science and Engineering (ICISE2009), pp.
493-496,
Nanjing
China.
Proceedings
Design and Analysis of Magnetic MEMS

Accelerometer for Inertial Navigation
M. Ameer Umar Malik, Umer Izhar, Umar Shahbaz, Javaid Iqbal
Department of Mechatronics Engineering
College of Electrical and Mechanical Engineering, NUST Islamabad, Pakistan
ameerumar@yahoo.com, umerizhar@yahoo.com
Abstract- Accelerometer is an important sensor of inertial

navigation system. Focus of new developments is a low cost, small
size sensor with high performance that meets the requirements of
navigation. In this paper a multi wafer magnetic MEMS
accelerometer design is proposed for inertial navigation.
Capacitive pick off is used to sense deflection of sensor mass and
permanent magnet & coil is used for rebalancing of deflected
mass of the sensor. Total two sides capacitive area of sensor is 18
mm2 with gap of 8m. Permanent magnet with thickness of 330
m and diameter of 700 m and soft magnet material coating is
used. Out of plane coil with total 64 turns, 8 layers and crosssection area 20m x20m is used. Fabrication process outlines
for both, the magnetic system and coils are proposed separately.
Stress analysis for flexures of sensing pendulum has been carried
out. Close loop frequency response analysis is carried out with a
proposed PID controller.
Key Words- Magnetic MEMS, Accelerometer
I. INTRODUCTION
Accelerometers are the important sensors of inertial
navigation
systems.
Traditional
high
accuracy
electromechanical accelerometers with quartz pendulum are
highly accurate with resolution of <1g and one year
repeatability of 50 g with input range of 50 g but also are
very expensive [1]. Whereas the best available MEMS devices
with input range of 10g have resolution of up to 250 g and
max repeatability of 25mg [2] but cost of MEMS sensors is
very low as compared to traditional electromechanical
accelerometers. Although accuracy of MEMS accelerometers
is yet not comparable to quartz flexure accelerometers but due
to their mass production process these provide very economical
solution particularly for low accuracy navigation applications.
MEMS accelerometers are widely being used in automobiles
but their accuracy is not enough for high accuracy inertial
navigation use. Variety of MEMS accelerometers are being
developed and reported for navigation applications. Most of
these are capacitive devices with electrostatic rebalancing
mechanisms. Recently reported ultra precision MEMS
accelerometer developed by Colibrys is also a capacitive
device working in close loop based on electro static
rebalancing [3-4].
In this paper a magnetic rebalancing mechanism for a
MEMS accelerometer is designed to get navigation grade
performance. Permanent magnets are being used in MEMS for
their repeatable performance. As the techniques of permanent

magnets fabrication in MEMS are being developed their use is
also increasing in high performance devices [5].
II. NAVIGATION GRADE ACCELEROMETERS
High performance accelerometers are required for high
accuracy navigation applications. Different performance
parameters can be specified for inertial navigation grade
accelerometers. In broad terms accelerometer performance can
be specified by two parameters i.e. bias and scale factor [1-6].
Accelerometers for navigation require high level of bias
stability, repeatability, temperature sensitivity and fine
resolution. Requirements of sensor parameters and comparison
for different requirements are discussed in [7]. Some of critical
parameters required for high accuracy navigation application
are given in table-1.
A variety of transduction mechanisms have been used in
MEMS accelerometers. Initially piezo-resistors were used in
MEMS accelerometers [7]. Then other principles like
capacitive, tunneling current, thermal, optical, electromagnetic
and piezoelectric etc are also utilized with their some
advantages and unavoidable disadvantages [8-9]. Capacitive
based sensing gives greater design flexibility than piezoelectric
quartz technology [6]. The capacitive devices are very popular
for their high precision.
TABLE:1 INERTIAL NAVIGATION REQUIREMENTS
Parameters
Values
Range
25 g
Bias Stability
50 g
Scale Factor Stability

Resolution
Frequency Range
Max. Shock in 1msec
Temperature Range
50 ppm
5g
DC-200 Hz
> 20 g
-40C to 80 C
III. PERMANENT MAGNETS IN MEMS

Use of electrostatic force is very common in
microelectronic systems due to its better force scaling at micro
scale level and also technology is highly developed as
compared to electromagnetic systems at micro level.[10]
Scaling differences of electrostatic and electromagnetic forces
222
is discussed in details by Edward P Furlani [5]. Magnetic field

from electromagnet unfavorably scales with diminishing sizing,
whereas the fields from permanent magnet are scale invariant.
Source of magnetic field in electromagnet is current but the
magnetic field in permanent magnet is residual magnetization
which arises at atomic level [5]. For actuation electromagnetic
is more stable for high force and for large gaps [11].
Use of permanent magnets was limited in MEMS due to
difficulties of their fabrication at micro scale level but now it is
increasing as different fabrication technologies are being
reported. Larger actuation distance can be achieved by using
magnets as compared to electrostatic actuation[5,12,13].
Permanent magnet is used in traditional navigation grade
electromechanical accelerometers. In MEMS accelerometer to
get g resolution and low noise of accelerometer a heavier
proof mass is recommended and to rebalance this proof mass
an electromagnetic system with a permanent magnet can be
used.
Proceedings
This can be modeled as a second order mass damping

system. Proof mass inertia is J, beams have effective stiffness
K and damping is D.
Transfer function is given as
H(s)
=
Where a is applied acceleration and is angular deflection
of proof mass.
IV. PROPOSED DESIGN

Accelerometer can be constructed and operated in open
loop or closed loop configuration. The basic construction of
accelerometer is such that a proof mass is suspended in case
and confined to a zero position by means of a spring, when the
acceleration is applied to the case of sensor, proof mass is
deflected with respect to its null position within the case, this
deflection of proof mass is proportional to specific applied
force along input axis. More accurate output can be obtained
by applying a rebalancing mechanism for deflected proof mass
[14]. This rebalancing mechanism can be either
electromagnetic
or
electrostatic.
In
traditional
electromechanical accelerometers the rebalance mechanism is
electromagnetic while in MEMS accelerometers it is
electrostatic.
A permanent magnet based MEMS accelerometer design is
proposed and analyzed. It is a three wafer design in which the
centre wafer with proof mass is sandwiched between two
symmetrical permanent magnet wafers. The bench mark of the
design is to use permanent magnet for rebalancing of proof
mass of MEMS sensor instead of electrostatic rebalancing.
Permanent magnet with volume of 0.7mm3 has been selected
to get bulk fabrication with powder size of 5~ 50 m and this
can be accommodated in 500 m thick wafer. Proof mass
includes the rebalance coils and capacitance area for pickoffs.
The designed proof mass has total area of 16 mm2 with
thickness of 500 m and coils mass is additional. This heavy
proof mass results in a low Brownian noise [15] and greater
stability and repeatability. The proof mass is suspended by two
flexure beams attached to the fixed frame of centre wafer. In
response to the applied input acceleration the proof mass
deflects and change in capacitance is detected. One side
capacitance area for pickoff is 9 mm2 which is large as
compared to in plane MEMS accelerometers, this result in fine
pickoff resolution. Rebalance force is applied by the passing
current through the coils present on the proof mass. This
current is the output of the accelerometer giving information
about the acceleration applied to the system.
Fig.1(a) Centre wafer with proof mass. (b) Three wafer device
For an applied input acceleration the total force depends on

the total proof mass of pendulum part. This input force is to be
balanced by coils. Lorentz force of the coils with current
flowing through is the rebalance force. The Lorentz force on a
conductor of length L and with the current i in an externally
generated magnetic flux density B perpendicular to the wire
is given by F = Li B. Rebalance force is directly proportional
to the magnetic field available in the working gap of coil. To
get maximum force magnetic field and coil placement need to
be optimized. Few analytical methods to get axial force
between a cylindrical magnet and a thick coil are discussed in
ref [16]. To get maximum magnetic field in working gap of
coil a magnetic loop is designed. Also to get maximum force,
magnetic field direction should be perpendicular to the current
flowing through the coil. Different configurations of magnetic
system with permanent magnet and soft magnet materials are
proposed, performance of each is evaluated to select the best
possible configuration. Magnetic flux density plot of selected
configuration is shown in in figure-2. Field variation in two
dimensions is simulated to get optimized location of coil.
Variation of flux density along z direction is plotted in figure-3.
The resolution of 10 g is selected as a target. Stiffness
hinges are designed so that the smallest input of 10 g
produces a deflection of 4.33e-7 rad. This minimum deflection
is used to estimate the overall gain required for pickoff system.
Overall pickoff gain of 2.32e6 is used and gain values of
different portions of pickoff system are calculated. Details are
discussed in closed loop analysis of system. All the parameter
values are calculated for this proposed design.
Stiffness - K= 2.32 g. mm / rad.
Damping - D = 1.16 g-s-mm.
Inertia - J = 0.0000307 g-mm-s2
223
Proceedings
Fig.4 Three Wafers of Magnetic MEMS accelerometer
Details and performance of this method is discussed in

details by Ololade D Oniku in ref [12]. Conventional
deposition of micro magnets with results is described in [13,
17]. Nd-Fe-B powder of size 15~ 50 m can be used to get the
size required in this design and this can provide remanence Br
of 0.5 ~ 0.7 Tesla. Doctors blade technique as discussed in ref
[12] can be easily adopted at wafer fabrication for this design.
Fig.2 Surface Plot of Magnetic Flux Density
V. FABRICATION PROCESSES
Design of this sensor system consists of three wafers as
shown in figure-4. Three wafers will be manufactured
separately and then bonded together to form a sensor. Two
wafers with permanent magnet are symmetrical. Coil is
fabricated on both sides of the third wafer. This third wafer is
to be fixed between two wafers with permanent magnet.
Fabrication processes for both magnet and proof mass coil are
proposed.
Permanent magnet is the key part in this design and fabrication
of permanent magnets has always been challenging in bulk
batch fabrication of MEMS. Permanent magnets are produced
by metallurgy process of mechanically pressing and heating.
But this process is yet not compatible with MEMS production
processes.
For fabrication of required size permanent magnet powder

fabrication is the best suited technique available. Fabrication
steps for complete magnetic system are proposed as
Cavity etching as per dimensions of magnet and
working clearance using DRIE process.
Coating of soft magnet material with thickness of 30 ~
40 m.
Fill sacrificial material and form a cavity for permanent
magnet powder.
Permanent magnet fabrication in cavity using powder.
Coating of soft magnet material with thickness of 30 ~
40 m above magnet.
Removal of sacrificial material to form working gap
for coil.
Fig.3 Magnetic Flux density along z direction
In MEMS conventionally permanent magnets are fabricated

by vapor deposition methods like sputtering, evaporation, pulse
laser deposition or by electrochemical deposition. By these
methods a layer of only few micrometers can be deposited.
Unconventional strategies like using magnetic powders of
5~50 micrometer sizes are being adopted to fabricate magnets
of required size at wafer levels. Permanent magnets fabrication
using magnetic powder and fabrication at wafer level is a
choice for this design.
Fig.5 Permanent Magnet Fabrication (a) Cavity etching for soft magnet
coating. (b) Coating of soft magnet material. (c) Cavity forming for
permanent magnet. (d) Permanent magnet fabrication using powder. (e)
Coating of soft magnet material above magnet. (f) Removal of sacrificial
material layer to form working gap.
Fabrication of rebalance coil in multi layers is required on

the centre wafer. Fabrications of spiral coils have been reported
and one such multi layer coil with 36 turns in two layers has
224
been reported by Chong H Ahn, and Mark G Allen [18]. The

rebalance torquer coil required in this design is of 64 turns with
8 layers. The spiral pattern is Cu electroplated in 8 layers with
each layer separated by an insulating layer.
Fabrication steps for coil are proposed as
Pattern planner spiral coil on wafer and electroplate
copper coil layer.
Sputter 1st insulating layer then pattern for 02 copper
links for 2nd layer and electroplate copper links.
Pattern for planner spiral 2nd layer on 1st insulating
layer and electroplate 2nd copper coil layer.
Sputter 2nd insulating layer then pattern for 02 copper
links (different location as compared to step b) for
3nd layer and electroplate copper links.
Pattern for planer spiral 3rd layer on 2nd insulating
layer and electroplate 3rd copper coil layer.
Repeat this process to complete 8 copper coil layers.
Etching of insulating layer.
This coil is to be fabricated on both sides of wafer. Two
coils are connected in series therefore via connection across
wafer thickness is also made. Capacitive area coating and
connection pads are connected through two flexure hinges.
Proceedings
VI. ANALYSIS
A. Structure Stress Analysis
Sensor structure can be subjected to different forces during
manufacturing, storage and operational life. Analysis of
designed structure is carried out to evaluate its survival/
performances in different conditions. Flexures are the critical
parts that are under stress, in free condition when centre part is
not fixed between two wafers the centre part can bend freely
and subjected to maximum stress as shown in figure-7. Input
moment is defined as M where P is the mass of proof mass
part of centre wafer.
=
as here are 2 flexure with section modulus S
B =
= 5.4012 Kg/ mm2
A =
= 6.3832 Kg/ mm2
Also deflection and angle of point B in free condition are

calculated.
B
+
+
)
)
= 0.0111 mm
= 0.0433 rad
Maximum deflection of point C in free condition is calculated

as
C
= B + *2L =
0.2494
mm
But in assembled form this maximum deflection is restricted by
the fixed wafers. Maximum deflection possible in assembled
form is the gap provided for capacitance pickoff. Stress
analysis in the condition when centre wafer is fixed between
two fixed wafers is also carried out.
Fig.7 Static condition when centre wafer is free
Fig.6 Coil Fabrication processes (a) Pattern planer spiral coil on wafer and
electroplate copper coil layer. (b) Sputter 1st insulating layer then
pattern for 02 copper links for 2nd layer and electroplate copper links.
(c) Pattern for planer spiral 2nd layer on 1st insulating layer and
electroplate 2nd copper coil layer. (d) Sputter 2nd insulating layer then
pattern for 02 copper links (different location as compare to step b) for
3nd layer and electroplate copper links. (e) Pattern for planer spiral 3rd
layer on 2nd insulating layer and electroplate 3rd copper coil layer. (f)
Repeat this process to complete 8 copper coil layers. (g) Etching of
insulation layer.
This condition is analyzed by using static equilibrium

conditions and general solutions for deflection of statically
determinate beams [19].
RA + RB = nP
--1
2L RB + M = nPL
--2
+(
) 2L = 0 ---
Moment is calculated by solving three equations and stress

due this moment M is calculated B = M/2S = 0.2381
225
Kg/mm2. This stress is produced by application of 1g input and

is much less than the stress calculated in free condition. From
this analysis it is concluded that in assembled form the flexures
can bear much higher inputs before failure.
Proceedings
sensor. Simulation analysis of the designed sensor is performed

to find out the behavior of sensor for critical performance
parameters. Measurement range is the minimum and maximum
acceleration that sensor can response along its input axis with
certain accuracy. As spring stiffness is designed lowest to get
minimum acceleration detection, in open loop operation the
small distance between the moving capacitance surface and the
fixed surface limits the maximum measurement range.
Performance of open loop accelerometer is very limited
concerning to measurement range, bandwidth and linearity [20].
In close loop operation maximum measurement range is
limited by the feedback force that can be applied to rebalance
the deflected mass.
This design can only be operated in closed loop with high
gain pick off system with negative feedback. To achieve
desired system dynamics PID controller coefficients should be
carefully designed [21]. Stiffness of flexure is very small to
detect smallest input. The deflection produced by smallest
input is to be detected by the pickoff system. Gains of pickoff
blocks are calculated to detect smallest input. Closed loop
analysis of the design is performed in Matlab. Architecture of
close loop with a PID controller HCONT is shown in figure-10.
Pick off gain and other parameter used for control and analysis
are HCA = 10 A/V. HPA = 6880 pF/Arc-min, KPSA =
0,2V/pF with PSA = 6.37e-6 sec. KTM =201 g-mm/A with
TM = 3e-5 sec. Steady state sensitivity of sensor depends on
mass of pendulum m, length of pendulum l and KTM of
rebalance torque system. As these parameters remain constant
for a sensor for this system steady state sensitivity if calculated
as Kss = ml / KTM = 0.1005 /201= 0.0005A.
Fig.8 Static equilibrium condition when centre wafer is fixed between two
wafers.
Stress analysis and deflection analysis is performed in FEM

software and the results are verified. Results of FEM analysis
for the free condition are shown in figure-9.
Mi
a
ml
n+
pF
Gp
HPA
(pF/arc- min)
HPSA
HCA
(V/pF)
(A/V)
i
HCONT
MTM
HTM Coil & Magnet
Fig.9 FEM stress analysis for free condition
B. Mechanical Noise
Intrinsic noise of sensor can be due to mechanical and
electronic components. Mechanicalthermal noise due to
Brownian motion is usually analyzed for MEMS sensor and is
expressed total noise equivalent acceleration (TNEA) as
Fig.10 Close loop architecture with output configuration
Bode Diagram
Gm = 75.5 dB (at 1.66e+004 rad/sec) , Pm = Inf
-50
C. Closed Loop System Analysis

Resolution, measurement range, bandwidth, non linearity
and noise are the parameters to be analyzed for performance of
226
-150
-200
-250
-300
-350
90
0
Phase (deg)
Where kB is Boltzmanns constant, T is absolute

temperature, w is natural frequency and Q is quality factor
of flexure. High mass and quality factors lead to low noise.
Noise level of this design is estimates as 4.3g/Hz.
Magnitude (dB)
-100
-90
-180
-270
-360
-450
2
10
10
10
10
10
Frequency (rad/sec)
Fig.11 frequency response plot of closed loop architecture
10
Frequency response analysis shows a bandwidth of 1.18e+4

rad/sec. Step response shows a rise time of 1.5154e-004 sec
with 27.5 % overshoot for this output configuration as is shown
in figure.12.
x 10
10 g
8g
1
6g
4g
0.5
Step Response
x 10
System: Gcsys
Peak amplitude: 0.000638
Overshoot (%): 27.5
At time (sec): 0.000434
Step Response- Displacement of pendulum
-3
1.5
Amplitude ( rad )
-4
Proceedings
2g
1g
0
0
0.005
0.01
Time (sec)
0.015
0.02
0.025
Amplitude
Fig.14 Deflection of sensor against step input
PID control parameters can be further optimized to achieve

best close loop performance of accelerometer.
0
0
VII. CONCLUSION
6
-3
Time (sec)
x 10
Fig.12 Step response of close loop architecture
Output response for a square input signal of 100 Hz is

shown in figure.
-4
Linear Simulation Results
x 10
6
Amplitude
In this paper a new magnetic MEMS accelerometer is

designed with capacitive pickoff using permanent magnet and
coil for rebalancing of proof mass instead of traditional
electrostatic rebalancing of MEMS accelerometers. Resolution
of 10 g can be achieved to fulfill navigation requirements.
Heavier proof mass results in reduced noise and also can be
rebalanced by electromagnetic force. Silicon structure provided
good thermal stability and magnets also provide low thermal
sensitivity as in traditional electromechanical accelerometers.
Permanent magnet can be fabricated in the bulk fabrication of
MEMS devices. Fabrication processes of permanent magnet
and coil are also proposed. In future study the device can be
fabricated and different performance parameters can be studied.
REFERENCES
-1
0.005
0.01
0.015
0.02
0.025
0.03
0.035
0.04
0.045
0.05
Time (sec)
[1]
Honeywell,
Q-Flex
http://www.inertialsensors.com
Fig.13 Response of close loop architecture
Maximum deflection of proof mass is limited by the

capacitance gap. In close loop operation the deflection of proof
mass must be minimum and less than the capacitance gap to
avoid electrical contact. To control the angular deflection the
close loop analysis is performed by transfer function as per
configuration is shown in figure 13.
Min
Gp
ml
HPSA
(V/pF)
HCA
HCONT
(A/V)
axis
analog
accelerometer
[3] Y.Dong, P.Zwahlen, A.M Nguyen,R. Frosio, F. Rudolf, Ultra

high precision MEMS accelerometer Transducers11, Beijing,
China,June5-9,2011.
[4] P. Zwahlen, Y Dong, A-M. Nguyen, F. Rudolf, J-M Stauffer,
P. Ullah, V. Ragot, Breakthrough in High Performance Inertial
Navigation Grade Sigma-Delta MEMS Accelerometer
[6] Ralph Hopkins, Joseph Miola, Roy Setterlund, Bruce Dow,

William Sawyer, The Silicon Oscillating Accelerometer: A HighPerformance MEMS Accelerometer for Precision Navigation and
Strategic Guidance Application Institute of Navigation 61st
Annual Meeting Cambridge, MA, June 27-29, 2005.
MTM
HTM
Colibrys
Single
http://www.colibrys.com
Accelerometers,
[5] E.P Furlani, Permanent Magnet and Electromechanical

Devices, Material analysis and Applications Academic press 2001.
Coil & Magnet
[2]
Navigation
HPA
(pF/arc- min)
pF
Fig.13 Close loop architecture for angular deflection
Deflection for a step input signal of different g is shown in

figure-14. Maximum deflection at input signal of 10 g is less
than 1.5e-3 rad. Minimum deflection of proof mass ensures
reduced non linearity of output.
227
[7] Navid Yazdi, Farrokh Ayazi, Khalil Najafi, Micro-machined

Inertial Sensors Proceedings of IEEE Vol. 86, No. 8 Aug 1998.
[8] Cenk Acar and Andrei M Shkel, Experimental evaluation and
comparative analysis of commercial variable-capacitance MEMS
accelerometers
[9] Teodor Lucian Grigorie The Matlab/Simulink Modeling and

Numerical Simulation of an Analog Capacitive MicroAccelerometer. Part1: Open loop MEMSTECH May 2008
Polyana UKRAINE
Proceedings
Optimizing the Geometry of an Electromagnetic Actuator IEEE

Transactions on Magnetics Vol. 48 No.9, September 2012.
[10] Qingkun Zhou, Azhar Iqbal, Pinhas Ben-Tzavi, Dapeng Fan,

Design, Analysis, and optimization of a magnetic micro
actuators Proceedings of the ASME 2009, November13-19
Florida ,USA.
[11] Tsung-Shune Chin, Permanent magnet films for application
in micro-electromechanical systems Journal of Magnetism and
Magnetic Materials 209 (2000) 75-79.
[17] Paul McGuiness David Jezersek, Spomenka Kobe, 100- thick Nd-Fe-B magnets for MEMS applications produced via a
low-temperature sintering route.
[18] Chong H. Ahn, Mark G. Allen. Micro- machined Planar
Inductors on Silicon Wafers for MEMS Applications. IEEE
Transactions on Industrial Electronics Vol. 45, No.6 December
1998.
[19] Andrew Pytel, Ferdinand L. Singer, Strength of Materials
[12] Ololade D Oniku, Benjamin J Bowers, Sheetal b Shetye,

Naigang Wang and David P Arnold Permanent magnet
microstructures using dry-pressed magnetic powders Journal of
Micromechanics and Microengineering published 12-june 2013.
[13] David P. Arnold., Naigang Wang., Permanent Magnet for
MEMS. Journal of Electromechanical Systems. Vol.18. No. 6,
December 2009.
[14] D.H.Titterton, J.L. Weston Strapdown inertial navigation
technology 1997 Peter Peregrinus Ltd.
[15] Ki-Ho Han , Young-Ho Cho, Self-Balanced NavigationGrade Capacitive Micro-accelerometer Using Branched Finger
Electrodes and Their Performance for Varying Sense Voltage and
Pressure Journal of Microelectromechanical systems, Vol. 12,
No.1, February 2003.
[16] Will Robertson, Ben Cazzolato, Anthony Zander, Axial
Force Between a Thick Coil and a Cylindrical Permanent Magnet:
228
[20] Teodor Lucian Grigorie The Matlab/Simulink Modeling and

Numerical Simulation of an Analog Capacitive MicroAccelerometer. Part2: Closed loop MEMSTECH May 2008
Polyana UKRAINE
[21] Yin Liang, Liu Xiaowei, Chen Weiping, Zhou Zhiping, High
resolution interface circuit for closed-loop accelerometer Journal
of semiconductor Vol.32, No.4 April 2011.
Proceedings
Feasibility Assessment of Running JP-8 Fuel in

Diesel Engine
M. Shahan1 , Ali Sarosh2,S. Javaid3
Department of Aerospace Engineering
Pakistan
1
heavy.cavalry.archer@gmail.com
2
alisarosh@cae.nust.edu.pk
3
saad.javaid@gmail.com
Abstract Single fuel concept (SFC) demonstrates the need for
a safe and cost effective fuel that can be utilized by light-duty as
well as specialized heavy-duty vehicles. Logistics and pipeline
operations are greatly simplified when SFC is implemented for
vehicles. The study considers the technical feasibility of using
aviation fuel JP-8 in diesel engines. It compares the effects on fuel
injection, combustion, engine performance and emissions when
using JP-8 fuel instead of diesel inside a compression ignition
engine. The performance of both fuels at various operating
conditions is calculated. Besides this, a detailed review of chemical
kinetics of diesel and JP-8 and the effect of thermodynamic
properties on engine performance and emission is also presented.
In addition, combustion analysis of both the fuels is done in which
endothermic enthalpies due to air and water at various fuel-air
ratios are calculated. The results reflect that use of JP-8 in diesel
engine leads to a penalty of torque and fuel consumption primarily
due to lower density and viscosity. The reduced cetane number
causes increase in ignition delay and premixed combustion of JP8. Modifications to engine are proposed so as to match the
performance of both the fuels. Analysis show that implementation
of SFC is best suited for supplication in heavy-duty diesel engine
vehicles
in ground battle field equipment leads to many advantages.

Logistics are simplified, oil lubricity is increased and exhaust
emissions are reduced. Single fuel concept also carried some
draw backs as well. Wear in fuel injection pumps exhaust valves
increases with the use of JP-8[4].Briefing charts regarding the
properties of JP-8 and other military fuels, different additives
that are used in fuels and their effect on fuel properties has been
presented by Schmitigal, Joel Tebbe and Jill in [5]. Similarly,
world jet fuel specifications have been given in [6]. A brief
review of the previous work done is this regard is given below:
Index Terms JP-8, Enthalpy, Cetane number, Ignition delay,

Pre-mixed combustion
INTRODUCTION
Single fuel policy states that Jet Propellant 8 (JP-8) fuel must
be used in all air and ground battle field vehicles. According to
a research, 38.6% of army logistics supply consists of fuel. JP-8
is considered as a safe fuel for storage as well as during
operations. The US army and NATO will be using JP-8 fuel in
their aircrafts and ground vehicles till 2025[1].Using JP-8 in
ground vehicles offered my advantages. It reduced nozzle
fouling, exhaust emissions, water entrainment and microbial
growth inside the fuel tanks and increased fuel filter replacement
intervals, storage stability and enhanced oil change and filter
replacement intervals[2]. JP-8 is a jet fuel and belongs to the
class of kerosene. The flash point of JP-8 is 100 degrees
Fahrenheit. It is specified by MIL-DTL-83133. It was first used
by the NATO countries having a NATO code F-34. JP-8 is
almost similar to Jet-A1 fuel except that it has three military
additives. Tests were conducted on standard diesel cycle to give
an overall idea of replacing JP-8 fuel with diesel [3].Using JP-8
229
Laboratory evaluation of JP-8 in diesel engine was

conducted as given in [7] in which JP-8 fuel was found
to be satisfactory for use in diesel engine but the
maximum engine power was reduced and leak in fuel
injection lines was observed. Compensation was done
by increasing thermal efficiency.
Engine testing with various blends of JP-8 was
conducted by April Covington with the perception that
in future an EPU will be designed to optimize engine
performance with JP-8[8]
Direct imaging and two color thermometry technique
was applied to verify the emission trends for JP-8 and
fossil diesel fuel in an optically accessible single
cylinder diesel engine. Image analysis focusing was
done to determine the characteristics of combustion.
Results verified that JP-8 emitted less smoke and
increased HC emissions [9].
Tests with petroleum, hydro processed and Fischer
Tropsch diesel and jet fuels were carried out in a direct
injection single cylinder diesel engine. Decrease in
ignition delay was quantified by increasing the derived
cetane number (DCN) of fuel. DCN decreased the
ignition delay[10]
Systematic study of ignition delay for different fuels in
2.44 L heavy duty single cylinder diesel engine was
conducted over a wide range of temperatures and
densities. Results indicated that fuel properties relating
to spray break have a nominal effect on the ignition
delay period [11].
Oxides of nitrogen are harmful to the environment.

During combustion, oxides of nitrogen are formed
depending on the temperature and pressure. Hence
there is a need to reduce nitrogen oxide emissions from
diesel engine. Control on NOx emissions was obtained
by introducing varying quantities of cetane improver
additive in diesel fuel by enhancing the ignition quality
of fuel[12]
Mie-scattering technique is used in [13] to determine
the spray characteristics of fuels which includes spray
angle and spray tip penetration. Spray characteristics
are in turn correlated to combustion and emission
trends of fuels
Tests conducted on S60 engine revealed that JP-8
reduced nitrogen oxide and particulate matter
signatures in the exhaust emission and hence it is
environmentally friendly. Torque and fuel economy
penalty were compensated by increasing the pulse
width of fuel [1].
Tests results carried out on 60 kW direct injection
diesel engine running alternatively on diesel, JP-8 and
JP-8 treated with varying quantities 2-ethylhexyl
nitrate showed that ignition delay decreases with the
addition of cetane improver additives [14]. Sound
analysis regarding maximum heat release rate, engine
efficiency, cylinder pressure and exhaust were also
presented.
Analysis of the usage of blends of biodiesel and JP-8 in
diesel engine as given in [15] revealed that specific fuel
consumption decreases and NOx emissions increase by
increasing the quantity of biodiesel in test fuels. It was
concluded that these fuel blends can be effectively used
inside diesel engine.
METHODOLOGY
The basic aim was the verification of the results whether JP8 could be used as a replacement of diesel fuel in diesel engine.
Main objective was to understand the effects on combustion and
performance of engine when replacing diesel with JP-8 in diesel
engine. The approach which was followed is as described:
Analysis of actual diesel and JP-8 fuel properties was
done so that effect of each property on engine
performance can easily be quantified
Heat of combustion analysis of both fuels was
performed to obtain an indirect insight about the
thermodynamic parameters of fuels and power output
of engine when replacing JP-8 with diesel fuel
Dynamometer testing of engine is done to determine
the performance parameters such as torque and power
output
Assessment of fuel economy and strategies for
matching the power output with JP-8 with different
adjustments
Proceedings
8.It relates to the fuel pressure and dynamic start of fuel injection
in the engine[16]
Cetane number provides a measure of ignition delay of fuel.
This property indicates the quantity of straight chain
hydrocarbons in fuel. Lower cetane number implies lower
proportions of straight chain hydrocarbons and in turn longer
ignition delay[17].
Heating value of a fuel determines its energy content. Higher
heating value results in higher power output of engine [11]. JP8 has higher heating value than diesel but its lower density leads
to lower heating value on volume basis.
Viscosity provides a measure of resistance to flow. Working
of fuel injection pump is directly related to viscosity of fuel.
Pump wear and leakage in injection pumps increases due to
lower viscosity. Generation of spray pattern is also governed by
the viscosity of fuel[18].
Volatility of fuel is governed by distillation curve. Heavier
fuel results in incomplete vaporization and combustion and it
causes soot and smoke quantity to be increased[18].
Hydrocarbons having multiple benzene rings are called Poly
Aromatic Hydrocarbons (PAHs). Diesel fuel contains higher
proportions of aromatics. Aromatics effect density, cetane
number and soot formation characteristics of fuel [17, 19, 20].
Environmental Protection Agency (EPA) requires that sulfur
content in the fuel should be as low as possible. JP-8 has lower
sulfur content as compared to diesel fuel so it is environmentally
friendly as compared to diesel.
CHEMICAL TESTING OF FUELS
Samples of JP-8 and diesel were taken and complete
chemical testing was done so that engine performance trends can
be co-related to chemical properties of fuels. Following results
were obtained from the chemical testing of JP-8.
Chemical testing of JP-8
FUEL PROPERTIES
Density effects the composition of fuel. Greater density
indicates lower volatility. Density of diesel is greater than JP-
230
Table 1 Test tests of JP-8

Test
Method
Test
Limits
Test
Result
ASTM D156
Report
+29
ASTM
D3242
0.015
max
0.007
Sulfur, Total % mass
ASTM
D4294
0.3 max
0.006
Sulfur, Mercaptan, %
mass
ASTM
D3227
0.003
max
0.0003
Volatility
Initial boiling point, C
10% vol. Recovery, C
ASTM D86
ASTM D86
ASTM D86
Report
205 max
Report
146.9
166.6
174.1
Appearance
Color, Saybolt
Composition
Total Acidity,
KOH/gm
mg

End point, C
Residue, % Vol
Loss, % vol
Flash point, C
Density @ 15C,
kg/3
ASTM D86
ASTM D86
ASTM D86
ASTM D86
ASTM D86
IP 170
ASTM 1298
Report
Report
300 max
1.5 max
1.5 max
38 min
775 to
840
188.6
214.9
228.9
0.5
0.5
40
802.2
Specific
Gravity
@15.6/15.6 C
ASTM
D1296
Report
0.8025
ASTM
D2386
-47 max
-50
Viscosity @ -20C, cst
ASTM D445
8 max
3.574
Combustion
Hydrogen Contents, %
mass
ASTM
D3343
Report
13.62
Smoke point, mm
ASTM
D1322
19 min
23
Specific Energy net,

MJ/kg
ASTM
D3338
42.8min
43.13
Naphthalene, % vol
ASTM
D1840
3 max
1.66
Contamination
Existent
gum,
mg/100ml
ASTM D381
7 max
Particulate
Contamination, mg/L
ASTM
D6452
1 max
0.8
Fluidity
Freezing point, C
Chemical testing of diesel

Table 2 Test results of diesel
Test
Appearance
Color
Composition
Total Acidity, mg
KOH/gm
Sulfur, Total % wt.
Volatility
Flash point, C
Specific Gravity @
15.6/15.6 C
Fluidity
Pour point, C
Method
Test
Limits
Results
ASTM
D1500
3 max
2.5
D974
0.5 max
0.01
D129
1.0 max
0.08
D93
ASTM
D1298
66 max
Report
60
0.833
ASTM
D97
-7 max
-7
Proceedings
Viscosity
@`20C, cst
Cloud point, C
Contamination
Ash content, %wt
Sediment, %wt.
ASTM
D445
ASTM
D2500
6.5 max
3.86
9 max
ASTM
D482
ASTM
D473
0.01
max
0.01
max
0.001
0.0008
COMBUSTION ANALYSIS OF FUELS

Combustion analysis of fuels was done on Chemical Kinetic
Analyzer to obtain an indirect insight about the thermodynamic
parameters of fuels and the power output of engine when
replacing JP-8 with diesel fuel
Figure 1 Chemical Kinetics Analyzer

The chemical kinetic analyzer works on the principle of heat
exchanging mechanism. Fuel is allowed to burn inside the
combustion chamber. Air is supplied to aid in the combustion of
fuel. Water flows through the outer walls of the combustion
chamber. When fuel burns, heat of combustion is absorbed by
air and water. In this way, we can calculate the heat released by
the fuel and the efficiency of combustion. We can also control
the amount of fuel and air that flows inside the combustion
chamber so that we can analyze the combustion efficiency at
different air to fuel ratios.
231
Proceedings
the equivalence ratio. However, efficiency increase for JP-8 is

higher than diesel
DYNAMOMETER TESTING
The engine used in experiment is ISUZU Truck 6BD1 102
kW engine. Complete specifications of engine are given in table
Table 3 Engine Specifications
Cylinder
6
Displacement
5.785L
Bore (mm)
102
Stroke (mm)
118
Compression ratio
16.5:1
Advertised Power/Speed
102/3000
(kW/rpm)
Max torque/Speed
402/1600-1800
(Nm/rpm)
Idle speed (rpm)
750
Ignition Sequence
1-5-3-6-2-4
Net Weight (kg)
555
Intake Valve Open at
18 (B.T.D.C.)
Intake Valve Close at
46 (A.B.D.C.)
Exhaust Valve Open at
48 (B.B.D.C.)
Exhaust Valve Close at
16 (A.T.D.C)
Figure 2 Schematic Diagram of Chemical Kinetics

Analyzer
Test Results
The dynamometer system used is capable to measuring

engine torque, RPM, power, fuel flow rate and oil pressure
inside the engine
Figure 3 Heat of Combustion of Diesel and JP-8
Figure 4 Combustion Efficiency of Diesel and JP-8

Comparison of Heat of Combustion and Combustion
Efficiency of Diesel and JP-8 is given in Figure 3 and Figure 4.
It clearly indicates that when JP-8 is used in Diesel engine, there
will be a compromise in power output. Heat release profile for
diesel is greater than JP-8 by about 5.6% and hence diesel engine
will give less power output when replaced with JP-8.
Combustion efficiency of both the fuels increases by increasing
Figure 5 Schematic Diagram of Dynamometer Testing

(1) Engine, (2) Dynamometer, (3) Shaft, (4) Flywheel, (5)
Exhaust Pipe, (6) Control Unit, (7) Fuel Measurement System
Density of JP-8 is less as compared to diesel and engine is
fueled on volume basis instead of mass. The compensation for
density can be done by increasing the duration of injection of JP8 so that same mass of both the fuels is injected inside the
combustion chamber and hence its power output can be
232
Proceedings
increased. Viscosity and volatility also affects the performance

of fuel injector. At the same time some adjustments can be made
to cater for the lower cetane number of JP-8 by changing the
injection timing. In this way, full understanding of the effect of
fuel properties on engine performance can be understood.
Correlation of performance trends with test results is done in
order to explain them and devise strategies for matching engine
output of JP-8 to that of diesel fuel.
Speed-Load combinations used during the testing of engine
are as follows:
1250 rpm, 40% load (low rpm - low load)
1600 rpm, 50% load (mid rpm - mid load)
Figure 8 Brake Specific Fuel Consumption @1250 RPM
and 40% Load
Testing at 1250 rpm, 40% load (low rpm - low load)
Lower density of JP-8 and leakage loses in the fuel injection

pump are two significant factors relating to the less fueling of
JP-8. Initially the fuel flow rate of JP-8 is 12% less than baseline
JP-8 and then it is increased by increasing the pulse width. After
the modification the difference is reduced to 5.4%. Similarly the
reduction in torque is due to lower fueling rate of JP-8 in spite
of the fact that JP-8 has a higher heating value than diesel on
mass basis. The increase in brake specific fuel consumption
(BSFC) is mainly due to increased frictional and mechanical
loses while running the diesel engine on JP-8.
The reduction in fueling rate can be compensated by
increasing the duration of injection of JP-8 so that same mass of
both the fuels is injected. The results of increased pulse width
are then compared with baseline diesel fuel.
Figure 6 Fueling Rate @1250 RPM and 40% Load
Similarly the quality of combustion can be improved by

advancement in injection timing of JP-8 to cater for its lower
cetane number. Combustion starts earlier in the cycle.
Consequent result is the reduction in mechanical loses and
increased brake power.
The effect of increasing pulse width and injection timing is
the increased torque and consequently higher power output.
BSFC is reduced when injection pulse width is increased. In this
way, performance of JP-8 can be match with baseline diesel fuel.
Table 4 Comparison of Brake power and BSFC
%
Difference
with
baseline
Diesel
5%
252.99
241.34
0.38%
JP-8
Increased
pulse
width
Brake
power
16
11
BSFC
242.30
7
309.9
5
Figure 7 Brake Power @1250 RPM and 40% Load
233
14.5
Increased
pulse
width +
crank
angle
15.2
Diesel
Testing at 1600 rpm, 50% load (mid rpm - mid load)
Proceedings
Table 5 Brake power and BSFC at 1600 RPM
Fueling
rate(kg/s)
Brake power
(kW)
BSFC
(g/kWh)
Figure 9 Fueling Rate @1600 RPM and 50% Load
Diesel
JP-8
%
Difference
0.002545
0.002775
-9%
41
41.8
-1.91%
223.503
237.931
-6.45%
At 1600 RPM, adjustment of pulse width is not required

because fueling rate is already higher due to increased load on
the fuel injector and higher pressure buildup. Hence the power
output of JP-8 is very much the same as that of diesel fuel. Brake
Specific Fuel Consumption of JP-8 is higher as compared to
diesel attributed to the lower cetane number of JP-8 and
increased mechanical loses as evident from figure above. But at
1600 RPM the temperature of engine is higher when compared
with diesel because of higher heating value of JP-8 and frictional
loses inside the engine.
RESULTS AND CONCLUSION
Figure 10 Brake Power @1600 RPM and 50% Load
Figure 11 Brake Specific Fuel Consumption (B.S.F.C.)

@1600 RPM and 50% Load
The basic aim is to the study of feasibility whether JP-8 fuel

can be used in diesel engine and the effects on engine
performance and fuel combustion. The methodology adopted
explains the rationale behind the differences in performance
achieved by two different fuels. The results of this experimental
study can be summarized as follows.
Properties of fuels differ from location to location due
to atmosphere and climate condition and the test limits
are set according to the environment where testing is to
be performed.
Properties of fuels have a major effect on the
performance parameters of engine like cetane number
governs the ignition delay period and lower viscosity
of JP-8 causes lower fueling rate, lower pressure build
up and leakage loses.
Although JP-8 has higher energy content as compared
to diesel but it results in reduction of power output
because of its lower density as the fuel is injected on
volume basis inside engine. This fact is verified by the
combustion analysis performed on Chemical Kinetics
Analyzer where diesel showed 5% greater heat output
When used inside the diesel engine, there is reduction
in power and torque output due to lower fueling rate of
JP-8 and lower viscosity. For example, in the above
case at 1200 rpm, fueling rate of diesel is 12% higher
than JP-8.
At low rpm, performance of JP-8 can be matched by
increasing the mass of fuel injected and advancing the
crank angle. The difference in power remained only
about 5% when the pulse width and crank angle
adjustment was applied.
At high rpm, there is no need for any engine
modification. JP-8 gives the same output as that of
234
Diesel because of higher pressure build up and load on

the fuel pump. But the engine heats up earlier because
of greater frictional loss.
Brake Specific Fuel Consumption for JP-8 is higher
than diesel fuel when used in diesel engine due to
higher frictional and mechanical loses and lower cetane
number of JP-8.
Analysis show that implementation of SFC is best
suited for supplication in heavy-duty diesel engine
vehicles
REFERENCES
[1]
G. Fernandes, J. Fuschetto, Z. Filipi, D. Assanis, and
H. McKee, "Impact of military JP-8 fuel on heavy-duty diesel
engine performance and emissions," Proceedings of the
Institution of Mechanical Engineers, Part D: Journal of
Automobile Engineering, vol. 221, pp. 957-970, 2007.
[2]
D. G. Weir, "Strategic Implications for a Single-Fuel
Concept," DTIC Document1996.
[3]
A. Command, "JP-8: the single fuel forward: an
information compendium," Warren, MI, 2001.
[4]
A. F. Montemayor, L. L. Stavinoha, S. J. Lestz, and M.
E. LePera, "Potential Benefits from the Use of JP-8 Fuel in
Military Ground Equipment," DTIC Document1989.
[5]
J. Schmitigal and J. Tebbe, "JP-8 and other military
fuels," DTIC Document2011.
[6]
E. Aviation, "World jet fuel specifications with avgas
supplement," Machelen: ExxonMobil Aviation, 2008.
[7]
W. E. Likos, E. C. Owens, and S. J. Lestz, "Laboratory
evaluation of Mil-T-83133 JP-8 fuel in army diesel engines,"
DTIC Document1988.
[8]
A. Covington, "The investigation of combustion and
emissions of JP8 fuel in an auxiliary power unit," 2011.
[9]
J. Lee, H. Oh, and C. Bae, "Combustion process of JP8 and fossil Diesel fuel in a heavy duty diesel engine using twocolor thermometry," Fuel, vol. 102, pp. 264-273, 2012.
[10]
S. Gowdagiri, X. M. Cesari, M. Huang, and M. A.
Oehlschlaeger, "A diesel engine study of conventional and
alternative diesel and jet fuels: Ignition and emissions
characteristics," Fuel, vol. 136, pp. 253-260, 2014.
[11]
D. A. Rothamer and L. Murphy, "Systematic study of
ignition delay for jet fuels and diesel fuel in a heavy-duty diesel
Proceedings
engine," Proceedings of the Combustion Institute, vol. 34, pp.

3021-3029, 2013.
[12]
K. Velmurugan and S. Gowthamn, "Effect of
CETANE Improver Additives on Emissions."
[13]
J. Lee and C. Bae, "Application of JP-8 in a heavy duty
diesel engine," Fuel, vol. 90, pp. 1762-1770, 2011.
[14]
G. Labeckas, S. Slavinskas, and V. Vilutiene, "Effect
of the Cetane Number Improving Additive on Combustion,
Performance, and Emissions of a DI Diesel Engine Operating on
JP-8 Fuel," Journal of Energy Engineering, 2014.
[15]
A. Uyumaz, H. Solmaz, E. Ylmaz, H. Yamk, and S.
Polat, "Experimental examination of the effects of military
aviation fuel JP-8 and biodiesel fuel blends on the engine
performance, exhaust emissions and combustion in a direct
injection engine," Fuel Processing Technology, vol. 128, pp.
158-165, 2014.
[16]
D. Kouremenos, D. Hountalas, and A. Kouremenos,
"Experimental investigation of the effect of fuel composition on
the formation of pollutants in direct injection diesel engines,"
SAE Technical Paper 0148-7191, 1999.
[17]
Y. Kidoguchi, C. Yang, and K. Miwa, "Effects of fuel
properties on combustion and emission characteristics of a
direct-injection diesel engine," SAE Technical Paper 01487191, 2000.
[18]
T. J. Callahan, T. W. Ryan, L. G. Dodge, and J. A.
Schwalb, "Effects of fuel properties on diesel spray
characteristics," SAE Technical Paper1987.
[19]
M. Lapuerta, J. Hernandez, R. Ballesteros, and A.
Durn, "Composition and size of diesel particulate emissions
from a commercial European engine tested with present and
future fuels," Proceedings of the Institution of Mechanical
Engineers, Part D: Journal of Automobile Engineering, vol.
217, pp. 907-919, 2003.
[20]
T. L. Ullman, K. B. Spreen, and R. L. Mason, "Effects
of cetane number, cetane improver, aromatics, and oxygenates
on 1994 heavy-duty diesel engine emissions," SAE Technical
Paper 0148-7191, 1994.
235
Proceedings
Artificial Intelligence Robot
Muhammad Shaheer Sajid
Muhammad Mohid
Collage of earth and atmospheric science

University of the Punjab
Lahore Pakistan
Osman.geomatics@gmail.com
LCAS
Lahore Pakistan
shaheersajid15@gmail.com
LCAS
Lahore Pakistan
mmoohhiidd111@gmail.com
Abstract: A.I.R or Artificial intelligence robot is a machine
These functions hold immense importance for what we

hoped to achieve. The use of obstacle detectors and line
following functions actually put up a vast frame of use of the
A.I.R. Examples include, factories, Metro Politian [1] and
even for military purposes. The fire detectors and path
detector functions allow the robot to be used as a fire escaping
gadget, public safety, health care management system
respectively [3]. A.I.R can be operated automatically and
manually. Arduino is used to control A.I.R through laptop.
This robot is 60% recycled.
II. OBECTIVES
which can be operated in two modes: manual and automatic.

This machine is specially designed to suit the needs for both
domestic and industrial purposes. A.I.R can be control and
operated within 10m of radius. A.I.R can detect obstacles
within distance of 1.5m. Then it will search the path without
obstacles. For automatics mode, it has been programed in such
a way, that we can define a line for its path then it will be
follow this path throughout its journey. Through the
To build an artificial intelligence robot for the public safety,

road navigation, health care management system, fire
detection, industrial and military purpose. Primary objectives
were to connect this robot to laptop and get on the spot
information regarding to defense purposes.
concentration of smoke with in its domains, A.I.R can detect

the location where the smoke occurred. Its can transmit videos
to the receiver station through the camera installed on it. This
video can be investigated on a laptop. Manual control can be
III. METHODS AND METHODOLOGY
done with the help of receiver.

A. Remote Controller
Keywords: Control system, artificial intelligence, security

robots, and robotics.
I.INTRODUCTION
A.I.R can be operated through remote control .It has a receiver

with its paired transmitters, which work on the principle of
radio waves. A transmitter sends radio waves to the receiver
which are then decoded and further processed (see fig 1).
There are four channels of the receiver which are attached to
the motor deriver circuit that derives the motor (see fig.2 & 6).
The radio circuit is a four data transceiver circuit, 2 controls
for each motor: forward and backward. Hence increasing
control over the robot.
Now-a-days, Robotics has become one of the essential tools

aiming to make human lives much more compatible and
easier. Over the past years, many different mobile robots have
been created which directly affect the scientific community;
such as crawling and legged robot. This paper deals with a
wheel robot [1]. Similarly, A.I.R has specifications which too
aim for the same outcome. The basic functions of A.I.R
include; obstacle detection, path following, line following and
smoke/fire detectors. Multiple mobile robots have advances on
single mobile robot [2]. Hence we make it a multiple mobile
robot. Alongside, our robot has the ability to be controlled
directly by laptop and also by a remote controller. It may also
be viewed by a light weight wireless camera. At this moment,
it is worth mentioning that A.I.R, though small in size, has
been very efficiently to build, also making it a low cost
autonomous robot. This actually helps large companies to buy
these robots and get access to their numerous advantages and
applications, cheaply.
B. Laptop Connection
In order to connect this robot to the laptop we have used
Arduino. The transmitter has a simple 4 channel encoder
circuit (see fig.2) that is connected to a development board
called Arduino (see fig. 3). The Arduino is programmed in
such a way to control it via a computer (see fig.4). Hence the
robot is a computer controlled one.
236
Proceedings
Figure 3: Arduino connections used.
Figure 1: Remote Control transmitter
Figure 4: Connectivity of the Arduino with radio transmitter.

The programming is done using LabVIEW, visual
programming software (see fig. 5). The program firsts
establishes serial communication with Arduino. There are 8
pins of the Arduino board being used for the movement of the
robot: 2 for each motion type i. e. forward, backward, right
and left (see fig.5).These pins are attached to the 4 channels of
the transmitter. So when the program runs, the user can
control the motion of the robot. The receiver is a decoder with
a radio receiver module tuned to the frequency of the
transmitter. It has decoder which decodes the signal and gives
out 4 channels. The four channels are then fed into a motor
driver circuit, 2 channels for each motor i.e. forward and
backward control for each motor (see fig.5).
Figure 2: Remote controller transmitter
C. Line following/Path following

Figure 7 shows the line following circuit on board of A.I.R.
The path following is a simple mechanism of reflection over
237
Proceedings
9). When both sensors are on white surface they turn the
outputs of both the operational amplifiers high. These outputs
are attached to the forward motion control pins of the motor
driver. Hence both the motors move forward (see fig. 10).
bright and dull surfaces. It is based on the working of a light

dependent resistor (see fig.8) over such surfaces.
Figure 7: Line following circuit on board of A.I.R.

Table 1: Specifications of the Line Following Circuit
Specifications
Explanation
LM741*2
Light dependent
resistor*2
10 kilo ohms*4
10 kilo Ohms*2
Figure 5: Programming in LabVIEW to conduct the action on

robot with keyword commands.
Two blue LEDs

470 ohms
Operational amplifier
used as comparator
to sense light reflected
from floor
Resistor to set reference
voltage
Variable resistor to
adjust threshold level
To transmit light on the
floor
Resistor limit current
through LEDs
Figure 6: Radio receiver circuit on the A.I.R.

Table 1 shows the table sheet used for line following sensor.
The robot has a pair of light dependent resistors which control
the motors independently (i.e. one connected to one motor).
The circuit which controls it is an operational amplifier used
as a comparator (see fig.8).The comparator compares voltage
of the light dependent resistor with the reference voltage
across it. When light falls on the light dependent resistor the
output of the amplifier turns high. There are two such circuits
each for one motor. The outputs are fed to the forward driving
input of the motor driver. The path detector has made circular
in shape and it is making of white color so that when A.I.R
moves over the surface it can detect its specific path (see fig
Figure 8: Line following circuit with L.E.D.s and Motor

drivers used in A.I.R.
238
Proceedings
4
D. Obstacle detection
Obstacle detection is the procedure of detection of obstacles in
the path of robot (see fig.12).Obstacle detection is based on
transmission and reflection of infrared rays. The infrared rays
fall on an object and reflect back (see fig. 13, 14 & 15). The
sensor receives the reflected rays and turns its output high or
low. This is used to drive the transistor which energizes the
relay. The relay acts as a switch and breaks the positive
connection of the line following circuit (see fig.13). The
circuit turns off and the wheels stop. Hence the AIR stops
when an obstacle comes in front of it. A.I.R detects the
obstacle within 1 meter of range and automatically stops 20cm
before the obstacle. The transmitter circuit is a dual oscillator
to oscillate first on infrared LED at 38 kHz and then oscillate
at 38 kHz frequency. This is so because the receiver used is a
38 kHz receiver and requires a dual oscillator to work. The
receiver circuit comprises of a 3 pin infrared receiver. When it
receives infrared rays it turns its output pin high and after
amplified by a transistor a relay is energized.
Figure 9: Line detector on the A.I.R.
Table 2: Specifications of transmitter for Obstacle detection
Figure 10: Mechanism of line following in forward motion

When one of the light dependent resistor fall on black surface
its resistance increases, this sets the output of the
corresponding amplifier low. The operational amplifier
attached to the forward pin of the corresponding motor shuts
down the motor.
Figure 11: Mechanism of the line following during turn.

Hence that motor is off while the other is on, this caused the
A.I.R to turn, in the direction to the off motor so that the both
the sensors are back on the white surface (see fig.11).
239
Specifications
Explanation
555 timers *2
Integrated circuit configured as A

stable multi-vibrator
1 kilo ohms
Resistor to set frequency
1.5 kilo ohms
330 ohms
Resistor to limit current through

led
56 kilo ohms
220 kilo ohms
470 ohms
Resistor to drive the transistor
10 nF
Capacitor to set frequency of

oscillator
4.7 nF
Capacitor to set frequency of

oscillator
Infrared L.E.D
Transmitter
BC547
Transistor to drive the first

oscillator
Proceedings
Figure 14: Infrared Transmitter for obstacle detection used in

A.I.R
Figure 12: Obstacle detection sensor on board of A.I.R
Table 3: Specifications of the receiver for Obstacle detection

Specifications
Explanation
BC547
Transistor to drive relay
330 ohms
Resistor limit current through

infrared sensor
470 ohms
Resistor drive the transistor
Tsop1738
Infrared sensor module

Figure 15: Principle of working infrared sensor to detect
obstacle in its path.
E. Motor driver
The motor driver circuit is a circuit based on the motor driver
integrated circuit L293D. It is a quad-H bridge motor driver
integrated circuit that can control two D.C motors in both
directions independently (see fig.16).
Figure 13: Receiver circuit of obstacle detection sensor.
Figure 16: Electric motors used in the A.I.R
240
Proceedings
6
F.
The camera is basically operated by Bluetooth. An app is used

to transfer live video from one device to another. The app used
is an android based one called Camera Remote. The app is
installed on two devices. One is configured as a camera while
other is configured as the display.
Smoke sensor
A.I.R has mounted with a gas sensor (smoke sensor) to detect

any leakage of gas. These gases are particular to methane,
propane and L.P.G (see table 4). The sensor has radioactive
isotope of the element Americium which acts as the heating
element (see fig 17) for the gases .The change also appears to
be the output voltage. This voltage can be measured to
determine the concentration of gas or even drive an alarm
circuit (see fig.18).
Table 4: Smoke sensor specifications.

Specifications
Explanation
MQ-2
10 Kilo ohms
LPG, methane, propane,

carbon dioxide sensor
Resistor to adjust the
sensitivity of the gas sensor.
IV. APPLICATIONS OF A.I.R

Wireless live video streaming
Can be used in hospitals for transporting patients and
other equipment through a line following system
Obstacle sensor is used to detect obstacles by
vehicles
For spying purposes by several agencies.
By the military for making robot based weapons
Can be used in industries to transport materials.
For security purposes
RESULTS AND CONCLUSIONS
This paper mainly describes the mobile robot which has

achieved the mentioned objectives for this study. Line
tracking makes this robot as path specific machines. It
detects the obstacle with in 1.5m of its radius, the
magnitude of the obstacle detection is going to increase
within 1.5m of its radius, A.I.R will stop before the
obstacle and will be able to change path through laptop
commands. Sensing the smoke in its buffer zone is only
possible when these gas molecules come in contact with
the gas sensor. Due to available sized antenna, A.I.R can
be controlled automatically within 20-27 m of domain.
But in future, this range can be increased so that its area
of application increases.
Figure 17: Gas sensor used in the A.I.R
Figure 19: A.I.R used for the research.
Figure 18: Testing of gas sensor

G. Wireless camera
241
Proceedings
ACKNOWLEDGEMENT
We would like to acknowledge the contribution of
Residence Officer, Prof. Dr. Sajid Rashid Ahmad,
Principal of Collage of earth and atmospheric science,
University of the Punjab, Lahore Pakistan for kind
contribution for the plagiarism evaluation for this study.
Also we would like to express our deepest gratitude to
Dr.Najam Abbas, Sectary ICASE 2015, Institute of Space
technology, Islamabad Pakistan for his very kind
contribution in publication of this effort to the public.
REFERENCES
[1]
[2]
R. Chandra Kumar, M. Saddam Khan, D. Kumar, R.

Birua, S. Mondal and M. Kr.Parai, 'OBSTACLE
AVOIDING ROBOT A PROMISING
ONE', International Journal of Advanced Research
in Electrical, Electronics and Instrumentation
Engineering, vol. 2, no. 4, p. 1430, 2013.
S. Shan, 'Mobile Robots Based Intelligent Fire
Detection and Escaping System', JOURNAL OF
COMPUTERS, vol. 8, no. 5, p. 1298, 2013.
[3] D. Punetha, N. Kumar and V. Mehta, 'Development

and Applications of Line Following Robot Based
Health Care Management System', International
Journal of Advanced Research in Computer
Engineering & Technology, vol. 2, no. 8, p. 2446,
2013
242
Proceedings
An Intelligent Approach for Edge Detection in Noisy

Images using Fuzzy Logic
Izhar1,2, Fayyaz3
R. Muhammad1 and N. Ahmed1
Mechanical Engineering Department,

CECOS University of IT and Emerging Sciences,
Peshawar, Pakistan
2
Institute of Mechatronics Engineering,
Peshawar, Pakistan
1
izhar@cecos.edu.pk
Department of Mechanical Engineering,

Pakistan Institute of Engineering and Applied Sciences,
Islamabad, Pakistan
AbstractEdge detection has widespread applications in the

field of feature extraction, machine learning, computer vision,
pattern recognition, remote sensing and satellite image analysis.
In satellite images it detects various regions of interest for
different applications. Though it has useful applications, but it is
a complicated task and it becomes more arduous when it comes
to noisy images. This paper reports an intelligent approach for
edge detection in noisy images. The proposed edge detection
approach is based on fuzzy logic. The proposed algorithm
employs a 3 x 3 window mask and a fuzzy inference system. The
values acquired from the window mask are subjected to the fuzzy
rules for edge detection. The proposed technique is successfully
tested on noise free as well as on noisy images. The experimental
results are also compared with the reported established edge
detection techniques such as Laplacian of Gaussian (LOG),
Roberts, Prewitt, Sobel and previously proposed fuzzy based
algorithm. The proposed technique when tested on a grayscale
image having 24 dB salt and pepper noise, detected 106 false
edge pixels. However, in comparison the reported edge detection
techniques like Sobel, Prewitt, LOG and previously developed a
fuzzy based technique have detected 3678, 4435, 835 and 2852
false edge pixels, respectively. From the experimental results it is
evident that the proposed edge detection technique provides
better solution to the edge detection problems in noisy images.
Index Terms Edge Detection, Noisy Images, Fuzzy Logic
I. INTRODUCTION
Edges in an image are contours produced as a result of
abrupt or sudden alteration in any of the (multiple)
characteristics at pixel level. These changes could be observed
due to change in texture, color, shade or light absorption. These
characteristics could further lead in estimating the orientation,
size, depth and surface features in an image [1]. Edge detection
has numerous applications in the field of robotics, medical
image analysis, geographical science, pattern recognition, and
military technology [2]-[7] etc. It is often the case that images
embody high frequency noise or irrelevant data which inhibits
the detection of continuous edge points [8] since edge itself is a
composition of high frequency data. The noise produces false
flags, which often mislead the techniques for an edge.
Several algorithms are employed for the detection of edges

in images [9]-[14]. The motivation behind each technique is to
overcome the limitations in previous methodologies. The
conventional techniques like Canny, Sobel, Robert, Prewitt and
LOG [9]-[14] incorporate the use of spatial differential filters
utilising local gradient. These filters recognise an edge as an
abrupt change of grey scale pixel intensities. These techniques
are well established. These techniques process the data in
relatively short time and are computationally efficient.
However, the conventional techniques are susceptible to affect
generated by noise during edge detection.
Jiange and Bunke [15] proposed an approximation of scan
lines method for edge detection. The technique in comparison
to other segmentation techniques produced considerably good
results. Genming and Bouzong [16] proposed a 5x5 window
mask for the detection of edges based on a fixed threshold
level. However, their limitation was its inadaptability to
regions with varying greyscale due to fixed threshold point.
Latest techniques incorporate the use of artificial immune
system, artificial neural networks, genetic algorithms with
particle swarm optimization and ant colony optimization [17][19].
Fuzzy set theory is another technique that is employed for
edge detection [20]-[21]. The method performs logical and
mathematical reasoning based on approximations rather than
crisp values. The technique therefore significantly reduces the
complexity of problems where fixed values cannot be
predicted. Kaur et al. [22] proposed a fuzzy logic based edge
detection technique employing a 22 window mask. Sixteen
fuzzy rules were defined for edge detection in the
algorithm .The results for edge detection was appreciable in
noise free images. However when subjected to noisy image the
technique detected the false edges produced due to noise as
well.
To address the short comings of the reported edge detection
techniques, a fuzzy logic based technique is proposed in this
work. The technique employs a 33 window mask guided by
fuzzy rule set for edge detection in noisy images. In our
previous work [23], we have used trapezoidal and gaussian
243
membership functions (MFs), however, in this work we have

preferred Pi and generalized bell shape MF to develop an
effective filter that is convenient to apply to both noise free and
noisy images. Moreover, algorithm developed in this work is
tested on noisy images having different noise levels.
II. PROPOSED METHODOLOGY
In Fig. 1 the conceptual framework of the proposed fuzzy
logic based edge detection is shown. The developed edge
detection technique for noisy images is based on fuzzy logic.
To extract the greyscale values of neighborhood pixels from
Proceedings
the input image, a 3x3 window mask is designed. The

grayscale values of the neighborhood pixels acquired from the
mask are pre-processed before applying to the fuzzy inference
system. A fuzzy inference system is designed that takes these
processed values as input. These values are subsequently
converted into the fuzzy plane. A fuzzy rule base is defined
that determine and show the edge pixels in the output image.
The output of the system is computed based on centroid
method and defuzzification is performed based on Mamdani
implication.
Fig. 1. Conceptual framework of the proposed edge detection approach.
A. Window Mask for Scanning the Image

A 3x3 window mask is designed for scanning the image, in
the proposed approach as shown in Fig. 2. The mask takes the
grayscale values,
of eight neighbourhood pixels with the
central pixel,
as the out pixel.
Before applying to the fuzzy inference system, the
grayscale values obtained from the mask are pre-processed.
Fig. 3 shows the processed mask, where
for x = 1, 2, 3,,8.
B. Fuzzy Membership Functions

In fuzzy inference system, MFs play a key role. In the
fuzzy set fuzziness is measured using MFs as they are the key
constituents of the fuzzy set theory. Based on the nature of
problem the type and shape of the MF should carefully be
selected as they have effects on the fuzzy inference system. In
the developed edge detection technique, for the input data, Pi
MFs are used, because they show reasonably improved results
with minimum error [24]. Moreover for the output data,
244
Proceedings
generalized bell MFs are used because they are smooth and
non-zero at all points.
Fig. 3. MF plots (a) Pi, (b) Generalized Bell.
C. Fuzzy Sets
Each input,
to fuzzy inference system is divided into
two fuzzy sets, Less and More. The output (pixel), from the
fuzzy inference system is also divided into two fuzzy sets,
Non-Edge and Edge. In Figs. 4 and 5, the associated MFs with
the input and output fuzzy sets are shown respectively.
Fig. 2. window mask for scanning the image.
Equation (1) represents the standard Pi MF.
)
(
)
1
(
(
)
)
Fig. 4. MFs of the input variable Pj.
where , ,
and are the various parameters of Pi MF, and
its details are depicted in Fig. 3(a).
Equation (2) represents the generalized bell MF.
2
|
where k, l, and m are the different parameters of the

generalized bell MF and its details are shown in Fig. 3(b).
Fig. 5. MFs of the output pixel
245
Table I lists the various terminologies and parameters of

both the input and output fuzzy sets.
TABLE II. FUZZY KNOWLEDGE BASE FOR THE DEVELOPED EDGE DETECTION
TECHNIQUE
TABLE I. PARAMETERS AND TERMINOLOGIES OF INPUT AND OUTPUT FUZZY

SETS
Linguistic Variable
Less
More
Less
More
Less
More
Less
More
Less
More
Less
More
Less
More
Less
More
Non-Edge
Edge
Parameter
Fuzzy Input P1
[0 0 23 47]
[23 47 255 255]
Fuzzy Input P2
[0 0 23 47]
[23 47 255 255]
Fuzzy Input P3
[0 0 23 47]
[23 47 255 255]
Fuzzy Input P4
[0 0 23 47]
[23 47 255 255]
Fuzzy Input P5
[0 0 23 47]
[23 47 255 255]
Fuzzy Input P6
[0 0 23 47]
[23 47 255 255]
Fuzzy Input P7
[0 0 23 47]
[23 47 255 255]
Fuzzy Input P8
[0 0 23 47]
[23 47 255 255]
Fuzzy Output G0
[3 2 10]
[3 2 248]
Range
Pi MF
[0 255]
Pi MF
[0 255]
Statement
R1
If G1==More && G2==More && G8==Less then G0 ==

Edge
R2

Edge
R3

Edge
R4

Edge
R5

Edge
R6

Edge
R7

Edge
R8

Edge
R9

Edge
R10

Edge
R11

Edge
R12

Edge
Pi MF
[0 255]
Pi MF
[0 255]
Pi MF
[0 255]
Rules
MF Type
[0 255]
Pi MF
[0 255]
Pi MF
[0 255]
Pi MF
[0 255]
Generalized
Bell MF
Proceedings
D. Fuzzy Knowledge Base

Fuzzy knowledge base or rule base in fuzzy inference
system is a set of linguistic descriptions [25]. Fuzzy rule base
plays a key role in fuzzy inference system as it makes
conclusions related to classifying an input or stabilizing and
adjusting the output. Fuzzy rule base for the proposed edge
detection algorithm consists of the following linguistic
descriptions as listed in Table II.

The developed edge detection technique is tested on a number
of grayscale images including noise free and noisy images. In
noise free greyscale images the developed technique has
successfully detected all type of edges as shown in Fig. 6. In
Fig. 6 (a) grayscale image of size 185 x 232 pixels having five
different regions covered by four boundary lines is shown. As
shown in Fig. 6 (c), the proposed technique for edge detection
has detected these four boundary lines (edges) successfully.
Similarly, as shown in Fig. 6 (b) the proposed method has
successfully detected edges in the greyscale (flower) image.
Fig. 6. Tested images: (a) Image having four different regions, (b) Flower, (c)
Edge detection in image having four different regions, and (d) Edge detection
in flower image.
246
The developed edge detection technique has the advantage

of detecting edges in the noisy images as previously discussed
(in the introduction). This is verified by detecting edges in an
image having 24 dB salt and pepper noise. To compute the
noise level in an image through peak signal to noise ratio
(PSNR), first based on equation (3) the mean square error
(MSE) is computed [26]-[27].
[
Where
and
represents the input noise free and noisy
images respectively. While and indicates the total number
of rows and columns of the input images respectively. Finally
Proceedings
the expression for the computation of noise level becomes as in

(4).
[
Where
denotes the maximum possible intensity value of
the pixel in the input image. The value of
for eight bit
unsigned integer data type image is 255.
The developed edge detection technique is applied to a size
500 x 232 pixels image having salt and pepper noise at a level
of 24 dB. The simulation results are compared with other
conventional and reported edge detection algorithms as shown
in Fig. 7.
Fig. 7. Comparison of experimental results in noisy image, (a) Original image, (b) Noisy image, (c) Sobel edge detection (d) Prewitt edge detection, (e) LoG edge
detection, (f) Previously developed fuzzy based edge detection technique [22], (g) The proposed method.
From the experimental results it is clear that the proposed

fuzzy based edge detection algorithm has detected a very few
false edge pixels in comparison to the other reported edge
detection techniques. The number of false edge pixels detected
by different reported edge detection techniques, when
subjected to images having various noise levels is shown in
Fig. 8. From Fig. 8 it is evident that the developed edge

detection technique when subject to a noisy image of 500 x 232
size and 24 dB noise level has detected 106 false edge pixels,
while other edge detection techniques for instance, Sobel,
Prewitt, LOG, and previously developed fuzzy logic technique,
247
after fine tuning their parameters have detected 3678, 4435,

835 and 2852 false edge pixels respectively.
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
Fig. 8. False edge detected pixels in a standard image of 500 x 232 pixels: A
comparision.
IV. CONCLUSION AND FUTURE WORK
[14]
This paper proposes and demonstrates a fuzzy logic based

edge detection algorithm for noisy images. The developed
technique employs a 33 mask guided by fuzzy rule set for
edge detection in noisy images. The developed technique has
successfully detected all the edge pixels in noise free and
noisy images. The developed algorithm is also compared with
other conventional and previously developed fuzzy logic
based edge detection techniques. The developed edge
detection algorithm when subjected to a 500 x 232 size
grayscale image having 24 dB salt and pepper noise has
detected a very few false edge pixels (106), while the reported
edge detection techniques like Sobel, Prewitt, LOG and
previously developed fuzzy logic have detected 3678, 4435,
835 and 2852 respectively. From the experimental results it is
obvious that in case of noisy images the proposed technique
provides better results.
In future an investigation on how to incorporate artificial
immune system with fuzzy logic to develop a hybrid
technique for edge detection is under consideration.
[2]
[3]
[4]
[5]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
[23]
REFERENCES
[1]
[15]
V. Torre and T. A. Poggio, On edge detection., IEEE Trans. Pattern

Anal. Mach. Intell., vol. 8, no. 2, pp. 14763, Feb. 1986.
G. O. B. Siliciano, L. Sciavicco, L. Villani, Modelling, Planning and
Control, in Robotics, Springer, 2010, pp. 415418.
Sirikan Chucherd, Edge detection of medical image processing using
vector field analysis, in 11th International Joint Conference on
Computer Science and Software Engineering (JCSSE), 2014, pp. 5863.
H. Lin, P. Du, C. Zhao, and N. Shu, Edge detection method of remote
sensing images based on mathematical morphology of multi-structure
elements, Chinese Geogr. Sci., vol. 14, no. 3, pp. 263268, Sep. 2004.
R. Rulaningtyas and K. Ain, Edge detection for brain tumor pattern
recognition, in International Conference on Instrumentation,
[24]
[25]
[26]
[27]
248
Proceedings
Communication, Information Technology, and Biomedical Engineering

2009, 2009, pp. 13.
A. A. Goshtasby, 2-D and 3-D Image Registration: for Medical, Remote
Sensing, and Industrial Applications. Wiley, 2005, pp. 3439.
A. Loft, K. E. Jensen, J. Lfgren, S. Daugaard, and M. M. Petersen,
PET/MRI for Preoperative Planning in Patients with Soft Tissue
Sarcoma: A Technical Report of Two Patients., Case Rep. Med., vol.
2013, p. 791078, Jan. 2013.
M. Setayesh, M. Zhang, and M. Johnston, Effects of static and dynamic
topologies in Particle Swarm Optimisation for edge detection in noisy
images, in 2012 IEEE Congress on Evolutionary Computation, 2012,
pp. 18.
J. Canny, A Computational Approach to Edge Detection, IEEE Trans.
Pattern Anal. Mach. Intell., vol. PAMI-8, no. 6, pp. 679698, Nov.
1986.
C. Y. H. X. Zhang Jin-Yu, Edge detection of images based on
improved Sobel operator and genetic algorithms, in 2009 International
Conference on Image Analysis and Signal Processing, 2009, pp. 3135.
A. Rosenfeld, The Max Roberts Operator is a Hueckel-Type Edge
Detector, IEEE Trans. Pattern Anal. Mach. Intell., vol. PAMI-3, no. 1,
pp. 101103, Jan. 1981.
R. A. Kirsch, Computer determination of the constituent structure of
biological images., Comput. Biomed. Res., vol. 4, no. 3, pp. 31528,
Jun. 1971.
L. Yang, X. Wu, D. Zhao, H. Li, and J. Zhai, An improved Prewitt
algorithm for edge detection based on noised image, in 2011 4th
International Congress on Image and Signal Processing, 2011, pp. 1197
1200.
F. Ulupinar and G. Medioni, Refining edges detected by a LoG
operator, in Proceedings CVPR 88: The Computer Society Conference
on Computer Vision and Pattern Recognition, 1988, pp. 202207.
X. Jiang and H. Bunke, Edge Detection in Range Images Based on
Scan Line Approximation, Comput. Vis. Image Underst., vol. 73, no. 2,
pp. 183199, Feb. 1999.
C. Genming and Y. Baozong, A new edge detector with thinning and
noise resisting abilities, J. Electron., vol. 6, no. 4, pp. 314319, Oct.
1989.
A. Naumenko, V. Lukin, and K. Egiazarian, SAR-image edge detection
using artificial neural network, in 2012 International Conference on
Mathematical Methods in Electromagnetic Theory, 2012, pp. 508512.
O. P. Verma and R. Sharma, An optimal edge detection using universal
law of gravity and ant colony algorithm, in 2011 World Congress on
Information and Communication Technologies, 2011, pp. 507511.
M. Setayesh, M. Zhang, and M. Johnston, Effects of static and dynamic
topologies in Particle Swarm Optimisation for edge detection in noisy
images, in 2012 IEEE Congress on Evolutionary Computation, 2012.
C.-C. L. Yau-Hwang , Chang-Shing Lee, A new fuzzy edge detection
method for image enhancement, in Proceedings of 6th International
Fuzzy Systems Conference, 1997, vol. 2, pp. 10691074.
S. E. El-Khamy, M. Lotfy, and N. El-Yamany, A modified fuzzy Sobel
edge detector, in Proceedings of the Seventeenth National Radio
Science Conference. 17th NRSC2000 (IEEE Cat. No.00EX396), 2000.
K. Kaur, V. Mutenja, and I. S. Gill, Fuzzy Logic Based Image Edge
Detection Algorithm in MATLAB, Int. J. Comput. Appl., vol. 1, no.
22, pp. 5760, Feb. 2010.
S. Mandal, J. Choudhury and S. Chaudhuri, In Search of Suitable
Fuzzy Membership Function in Prediction of Time Series Data, IJCSI
International Journal of Computer Science Issues, vol. 9, no.3, 2012.
Izhar, I. U. Haq, K. Shah, T. Khan, K. Azam, S. Anwar, Fuzzy Logic
Based Edge Detection for Noisy Images Technical Journal, UET
Taxila, vol. 20 (SII), no. II (S), 2015.
O. Karray, and C. De Silva, Soft computingand intelligent systems
design: theory, tools and applications, Pearson Addison Wesely, Essex
CM20 2JE, England, 2004.
F. Luisier, T. Blu and M. Unser Image Denoising in Mixed PoissonGaussian Noise, IEEE Transactions on Image Processing, vol. 20, no.
3, pp. 696-708, 2011.
J. Liu, Z. Huan and H. Huang, Image restoration under mixed noise
using globally convex segmentation, Journal of Visual Communication
and Image Representation, vol. 22, no.3, pp. 263-270, 2011.
Proceedings
A Survey of Active ITU-R P-Series Propagation

Models
Sundus Najib1, Ibrar Ali Khan2, Muhammad Waleed
Aftab1, Naveed Mufti1
Rafia Mehreen
Department of Telecommunication Engineering
University of Engineering and Technology, Taxila
Taxila, Pakistan
Department of Telecommunication Engineering

2
Department of Electrical Engineering
University of Engineering and Technology, Peshawar
Peshawar, Pakistan.
sundus.najib@gmail.com
Abstract ITU-R (International Telecommunication Union)

P-series (Propagation-series) recommendations encapsulate the
propagation models and data vital for modeling of any radio
communication path, including space and terrestrial paths. For a
new researcher, an understanding of these recommendations
prior to their application in various scenarios is desirable. This
paper presents an introduction to propagation, radio channel
impairments and previous review studies, followed by review of
thirty-one
recommendations
applicable
to
numerous
indoor/outdoor propagation scenarios. It is observed that some
outdoor propagation models can be used for partial indoor
scenarios while the only indoor model provided by ITU-R series
can also be used for various outdoor setups. Certain
recommendations
are
interconnected
with
other
recommendations, developing a need for their comprehensive
study.
Keywords ITU-R; Radiowave Propagation; Indoor/Outdoor
propagation models; P-Series.
I. INTRODUCTION
The International Telecommunication Union (ITU)s
Radiocommunication Sector primarily works for global
harmonization of the use of radio frequency spectrum. In
addition to supervising the international radio spectrum
regulation and satellite orbits resources, ITU-R also develops
Radiocommunication standards and recommendations for the
efficient use of spectrum in Radiocommunication systems [1].
Radio spectrum has a vast range of applications in
considerable operational environments. Three main
mechanisms namely refraction, diffraction and scattering
make it possible for the radio waves to propagate beyond the
horizon in terrestrial communication [2]. Certain factors can
have different effects along the path of transmission, resulting
in the loss of information signal. For communication to
succeed, modeling of the radio spectrum/system to overcome
and minimize these channel impairments is vital [3].
Member States and Industry and ultimately converting

different inputs into ITU-R Recommendations. It takes years
of research and hard work for developing and modifying each
recommendation, thus making it valid and reliable. The
Recommendations are further classified on the basis of
services [1]. There are currently seventy-eight active
recommendations dealing with Radiowave Propagation [4].
Each recommendation is valid for specific conditions and
frequency bands. At the time of writing, there is no known
comparison/review of the ITU-R Recommendations pertaining
to propagation. It is felt that, for a new researcher, it becomes
difficult to understand, sort and pin point the most relevant
ITU-R propagation model for his/her modeling/
scenario/investigation. This review paper is aimed at
providing a comparative summary of the propagation models
contained within the ITU-R P-series Recommendations. It is
hoped that this would help the researchers to get an overview
of the propagation models within the P-series and enable them
to select the most relevant recommendation for modeling their
specific scenario.
The order of remaining paper is as follows: Section-II
explains basic propagation mechanisms, followed by radio
channel impairments. Section III discusses previous review
studies while in Section-IV, ITU-R indoor and outdoor radio
propagation models are discussed. Finally Section-V and
Section-VI puts forward the guidelines and conclusions along
with future work of this review paper.
II. PROPAGATION MECHANISMS AND RADIO CHANNEL
IMPAIRMENTS
A. Propagation Mechanisms
Many modeling techniques for radio channel modeling

have been put forth by researchers from time to time for
effective radio communication. ITU-R encourages research
and
investigations
into
different
aspects
of
radiocommunication by proposing questions (ITU-R
Questions), constituting Study Groups, seeking input from
249
1) Refraction:
In radio propagation, the atmospheric bending of the radio
signal away from the straight path is termed as atmospheric
refraction [2]. A drop in temperature and pressure with
increasing altitudes results in decreasing value of refractive
index with height, contributing to bending of the wave
towards the ground. Atmospheric refraction is calculated in
units of N [2]:
Where, is the refractive index.
Proceedings
Where,
Ps= Signal power at transmitter (Watt)
Pd=Signal Power at Receiver (Watt)
2) Free Space Loss (FSL):
FSL is assumed for simplification purposes. A signal can
be visualized as spreading out from a transmitter. As it travels
away from the source, it spreads out in the form of a sphere.
The surface area of the sphere increases as distance from
transmitter increases. Following the law of the conservation of
energy, the signal intensity at a point must decrease as the
surface area of the sphere increases [2].
Fig. 1.
Refraction Types [2]
In order to follow the earths curvature, the change in

refractivity with height, termed as vertical refractivity
gradient, denoted by dN/dh, should be equal to -157 N
units/kilometer [2]. When the gradient exceeds -157 Nunits/kilometer, trapping of signals inside ducts can occur. A
less negative refractivity gradient causes signals to be superrefracted while gradients more positive than -40 Nunits/kilometer can cause sub-refraction. Fig-1 shows different
refractive types and associated values of gradients.
2) Diffraction:
Diffraction occurs due to the obstructions present in the
path of transmitter and receiver antennas [3]. Objects with
sharp irregularities and impenetrable bodies such as buildings,
vegetation, terrain etc. are the main reason for the diffraction
of radio signal. Three common procedures for diffraction
modeling are Bullington method, Epstein-Peterson method
and Deygout (also known as the principle edge method) [2].
Bullington method is widely used in ITU-R P-Series
Propagation Models for diffraction calculations which is based
on constructing a comparable knife edge at the junction of
transmitter/receiver horizons.
3) Scattering:
When radio signal encounters objects whose size are of the
order or less than that of propagating wave, it gets scattered
[3]. For long range Very High Frequency (VHF)/Ultra High
Frequency (UHF) propagation, tropospheric scattering plays
an important role, where signals are scattered to the receiver
located beyond the horizon. A limit to the propagation range is
impinged by the joint bulk made by the overlap of
transmitter/receiver antenna patterns which needs to be in the
troposphere [2].
B. Radio Channel Impairments
1) Attenuation:
Attenuation can be defined as the gradual degradation of
signal power during its transmission over a radio path [5]. It is
measured in units of decibels (dB). Attenuated power can be
calculated by:
250
3) Multipath Effects:
Due to the irregularities and non-uniform behavior of the
radio channel along with the straight path in which the signal
is transmitted, signal may reach the receiver by following
other directions. The resultant signal at the receiver is then the
combined effect of these multipath signals which either
interfere constructively or destructively, depending on the path
differences and antenna heights.
4) Fading:
Fading is a random process that may vary with position,
frequency and time. Fading encountered due to multipath
effects can be categorized as multipath fading while the
shadowing of radio signal from different objects cause
shadowing fading.
5) Signal Penetration into Buildings:
Radio signals are attenuated when they come in contact
with a building/hill etc. Shape of the building, frequency of
operation and materials of the building greatly affect the
penetration phenomena. Normally signals are more attenuated
by metals than plywood/concrete. Aluminum acts as a
reflector medium while gypsum and asbestos have higher
signal absorbing capabilities than other materials [6].
III. PREVIOUS REVIEW STUDIES
Sumit Joshi et al [7] and Govind Sati et al [8] present
reviews of some commonly used empirical, stochastic and
deterministic propagation models, namely FS path loss model,
Okumura model, COST 231 Hata model, Stanford University
Interim (SUI) Model, ECC-33 model (suitable for urban
environments) and COST-231 Walfish-Bertoni model for
small height buildings. Furthermore, Sumit Joshi et al [7]
provided a MATLAB simulated comparison of the models
they tested and generated an error factor to be added with
Bertonis model to achieve higher fidelity for scenarios under
their investigation. Pooja Rani et al also reviewed above
mentioned models along with two additional propagation
models i.e. Longley-Rice model and Two-Ray Ground model
[9].
A performance comparison of Longley-Rice Model, ITUR P.1546 terrestrial Model and HATA-Davidson Model using
Height Above Average Terrain (HAAT) for Digital Video
Broadcasting Television shows that ITU-R P.1546
recommendation is only reliable for distances shorter than 50

kilometers while the Longley-Rice model incorporated with
terrain information can lead to better radio modeling
performance but at the cost of computational efficiency [10].
Empirical and deterministic propagation models for terrestrial
communication are reviewed by Magdy F. Iskander et al [11]
and ray-tracing methods for enhanced deterministic models
are discussed thoroughly. Unlike other studies, Tapan
K.Sarkar et al [12] provide a more comprehensive review of
several statistical and site-specific models that shed light on
various scenarios of smallscale/ large-scale fading. The study
proposed a new numerical approach for prediction of
propagation mechanisms. For indoor environment, an advance
ray tracing method is put forward for locating the rays paths
while Finite Difference Time Domain (FDTD) method is used
for analyzing the propagation of radio wave through the walls
of building.
All the above referenced comparative studies focus on
similar indoor and outdoor propagation models. No review has
been conducted for discussing only ITU-R propagation models.
This presents an urge to address these models for convenience
of other researchers.
IV. DISCUSSION OF ITU-R INDOOR/OUTDOOR PROPAGATION
MODELS
At the time of writing, ITU-R Propagation Section has in
force seventy eight recommendations, related to indoor
propagation models, outdoor propagation models, and global
maps along with data, attenuation models, diffraction models,
noise model and list of other factors that should be considered
in modeling of radio paths [4]. Our review study focuses on the
Indoor and outdoor models. In this regard, a total of thirty-one
recommendations have been studied and reviewed in this
paper. The recommendations numbers along with area of
concern, frequency of operation and path length applicability
are listed in Table-1.
A. Indoor Propagation Model
Unlike many outdoor models, ITU-R has only published a
single Indoor Propagation model, as part of its P.1238
Recommendation [13]. The model, can however, be applied to
partial outdoor scenarios [14]. This model covers the
mechanisms and factors related to the radio propagation of
indoor signals operating in 900 MHz to 100 GHz frequency
range. The Path loss models are divided in two categories,
namely site-general (little site information needed) models and
site specific (detailed information of building) models. The
models assume that the portable user equipment and the base
station are located inside the building premises. Statistical and
deterministic delay spread models are also discussed. The
impact of polarization and antenna radiation patterns is
discussed for three different cases i.e. LOS case, obstructed
LOS case and cases where portable radio terminals are
randomly oriented resulting in the scatter of transmitted
polarization energy into orthogonal polarizations. Dependence
of signal strength and quality on location of the transmitter and
receiver, effect of buildings structure and its materials
permittivity, and the effects of motion of objects in room are
also thoroughly discussed in this recommendation.
251
Proceedings
B. Outdoor Propagation Models:

1) Aeronautical and Space Models
P.528 is an aeronautical and satellite communication based
model, calculating basic transmission loss [15] of aeronautical
and satellite services operating in frequency range of 12515500 MHz. It uses an interpolation method to determine
basic transmission loss data, from set of curves that are valid
for ground-satellite, ground-air, air-air, satellite-satellite and
air-satellite links. This transmission loss is predicted on the
basis of distance between antennas, the heights of the antennas
above mean sea level, the frequency, and the time percentage.
P.682 provides guidance on Earth -Space aeronautical mobile
telecommunication systems planning by incorporating models
to be used with P.618 [18]. The models therein take account of
effects generated by signal multipath and scattering from the
surface of the earth [16]. The propagation mechanisms
discussed in P.1409 [17] should be taken into consideration
when planning a radio communication system comprising of
stations at high altitudes and other higher level platforms in
stratosphere, and in the studies of sharing and compatibility.
Recommendations P.531 [19], P.618 [18], P.619 [20], P.679
[21], P.680 [22], P.681 [23] and P.840 [24] are developed for
planning space, satellite and earth-satellite systems. P.531
provides data and methods that help in satellite systems
planning by evaluating the ionospheres effects on Earth-space
paths at frequency range from 0.1 GHz to 12 GHz. P.618 also
contributes to the planning of earth-space communication
planning. P.619 puts forward model to be used for calculating
interference
between
earth
and
space
stations.
Recommendation P.679 uses methods from P.618 and
additional data to model broadcast satellite systems. P.680
also uses methods from P.618 and incorporates other data for
designing Earth-space nautical mobile satellite systems. P.681
deals with modeling of land mobile telecommunication
systems between Earth and space. P.840 incorporates the
models, graphs and global maps needed for the calculation of
attenuation caused fog and clouds in the atmosphere.
2) Terrestrial Models
Total of twenty terrestrial propagation models are
discussed in this section. P.452 [25] addresses the microwave
interference and propagation loss encountered between earth
stations operating at frequency range of 0.1 GHz to 50 GHz.
Recommendation P.525 concerns the free space planning and
provides guidance on the attenuation effects that should be
considered in designing a radio link that assumes free space
environment [26]. For evaluating the diffraction effects on the
received signal strength, P.526 recommends a number of
models that are applicable to various obstacle types and
variety of path geometries such as spherical earth surface and
irregular terrain [27]. P.530 [28] comprehensively describes
the propagation effects that should be tackled when planning a
fixed digital line of sight (LOS) link. ELF, VLF and LF
terrestrial radio waves are confined within the space between
the Earth and the ionosphere in order for them to travel to
greater distances. For this purpose, recommendation P.684
proposes methods for developing services in the frequency
range below 150 KHz [29]. P.842 [30] focuses on the issue of
reliability and compatibility in planning and design of High

Frequency (HF) radio communication systems. Considering
the fact that scattering from ionization caused by meteor trails
can provide a convenient way of communication at HF and
VHF, Recommendation P.843 [31] tackles with this issue of
meteor-burst propagation. P.844 provides information that
should be considered when designing radio communication
systems that rely on sharing of frequencies within the 30
MHz-3 GHz Range [32]. Planning of broadcasting services
falling in the LF and MF bands i.e. 150,000 Hz to 1700,000
Hz, having path lengths from 50 km to 12000 km, are
discussed in P.1147 [33]. P.1321 [34] shed light on the
characteristics of LF/MF ground-wave and sky-wave
propagation that might cause an effect on the digital
modulation techniques used in those bands. In designing and
planning of terrestrial land mobile and broadcasting services,
propagation characteristics whose impact cannot be ignored in
the proper modeling of these services are mentioned in P.1406
[35]. Recommendation P.1410 [36] addresses radio
communication systems operating in frequency ranges of
3GHz to 60 GHz (broadband services) and provides guidelines
for line-of-sight and alternative non-LOS propagation
mechanisms. P.1411 [14] is a model for short range
propagation in an outdoor scenario over the frequency
spectrum of 300 KHz to 100 GHz but it can be used as good
modeling technique for cases where some indoor factors are
also involved. A method constituting of interpolation and
extrapolation techniques from field strength curves that are
based on empirical data is presented in recommendation
P.1546 which can be applied to all the polarization types [37].
These curves are plotted as a function of path length, effective
height of the antenna, frequency of operation and percentage
time of the year. This method predicts the radio propagation
behavior in point-to-area scenarios for land based services
operational in the frequency band of 30,000 KHz to 300 MHz.
Recommendation P.1791 [38] provides procedures valid for
frequency ranges from 1 GHz to 10 GHz, to calculate path
loss for Ultra-wideband (UWB) indoor and outdoor
environments for both direct and obstructed paths, and
estimating the power received by a conventional narrow-band
receiver from an ultra-wideband transmitter. The method
presented in P.1812 [39] gives a thorough study of terrain for
point to area services which are land based and operate in the
frequencies between 30,000 KHz to 300 MHz and paths
starting from 0.25 kilometer and going up to 3000 km
distance, with both terminals having maximum height of 3 km
above ground. Radio communication can also be achieved by
using optical and infrared spectra. For taking advantage of this
resource, Recommendation P.1814 provides propagation
estimation procedures for terrestrial free-space optical systems
planning. Attenuation caused by the presence of fog in the
atmosphere, rain, snowfall and attenuation of signal during
normal sunny days is also addressed in this recommendation.
It also covers scintillation and impairments by sunlight [40].
P.1816 [41] is an urban/sub-urban propagation model that
guides on the forecasting of time based profiles and profiles of
spatial elements for broadband land mobile services operating
252
Proceedings
in the frequencies between 700 MHz to 9000 MHz and

distances 0.05 km-3 km. In 2013, ITU-R published a widerange terrestrial propagation model, P.2001 [42], that predicts
the Path Loss due to both enhancement of the radio signal and
occurrence of fading mechanism over the range from zero
percent to hundred percent of an averaged year and is
applicable to the frequency range from 30 MHz to 50 GHz.
Finally, P.2040 deals with the issue of impact of buildings
infrastructure on propagation of the radiowaves i.e. materials
types used in its construction and the shape of building [43].
V. GUIDELINES FOR RESEARCHERS
We recommend following procedure to be adopted for
selecting a proper recommendation prior to modeling a
scenario.
1. ITU presents two categories of recommendations i.e.
terrestrial and space models, each dealing with separate set of
environmental conditions. Thus foremost, it is important that
scenario should be distinguished whether it is a terrestrial or
space dominant environment.
2. The communication type must also be specified
whether it is point-to-point, point to area or indoor
communication scenario.
3. On the basis of working environment, select a
suitable recommendation that covers the range of operational
frequency and distance. It is of crucial importance to consider
frequency limitation because certain parameters in
recommendations are frequency dependent and if not taken
care of, can produce inaccurate results.
4. Shortlisted recommendations should then be assessed
in terms of scope of that recommendation, whether it tackles
with the desired metrics and purpose of study. For example, if
interference measurement is the required analysis then one
should consider interference modeling recommendations.
5. The final selected model should then be read in detail
and desired input parameters be collected.
6. Some
recommendations
rely
on
other
recommendations for calculating certain parameters. A
thorough study of referenced models is also needed.
Although recommendations provide step by step procedures
for calculating each parameter, they require intense
mathematical calculations and a good understanding of
complex mathematical problems.
VI. CONCLUSIONS AND FUTURE WORK
The propagation models and relevant procedures contained
within
the
ITU-R
Propagation Series
(P-Series)
Recommendations comprehensively address all the possible
radio propagation environments from indoor to outdoor space,
urban/suburban and aeronautical cases. A detailed
understanding of these recommendations prior to their use in
ones research is needed. Recommendations rely on each other
for the calculation of certain atmospheric and modeling
parameters. Certain short range outdoor models and the sole
indoor model available, can be used for partial indoor/outdoor
scenarios. Also each recommendation covers some specific
environment and frequency of operation, limiting its area of

use.
The latest ITU-R Propagation Model, P.2001, presents a

wide-range propagation model for general purpose terrestrial
communication. It covers frequency range from 30 MHz to 50
GHz and distances from 3 km to at least 1000 km. Unlike
most of the previous models, the model predicts path loss
effectively in the range from 0% to 100% of an average year.
The model can be compared with other relevant propagation
models for a study on a specific outdoor scenarios.
Researchers are encouraged to benefit from the

comparisons and information contained in this review in
making decisions about most suitable/applicable models for
the scenarios under investigation.
TABLE I.
S.No.
Proceedings
ITU INDOOR/OUTDOOR RECOMMENDATIONS SPECIFICATIONS [1,44]
Model type
Scope
Recommendation
Number
P.452
Terrestrial
Interference prediction and path loss estimation
2
3
4
P.525
P.526
P.528
FS attenuation calculation
Diffraction calculation
Transmission loss calculation/ path loss
P.530
Terrestrial
Terrestrial
Aeronautical/ satellite
services
Terrestrial
P.531
Satellite System
P.618
Earth-space model
8
9
P.619
P.679
Earth-space model
Satellite
10
P.680
Satellite
11
P.681
Satellite
12
P.682
Aeronautical
13
P.684
Terrestrial
ELF, VLF and LF terrestrial radio services
14
15
16
P.840
P.842
P.843
Earth-space model
Terrestrial
Terrestrial
17
18
P.844
P.1147
Terrestrial
Terrestrial
Attenuation due to clouds and fog

Reliability/compatibility prediction
Transmission loss, meteor flux variations, antenna and
frequency considerations in Meteor-burst communication
Ionospheric communication
Estimating sky-wave field strength in Broadcast services
19
P.1238
Terrestrial
20
21
22
23
P.1321
P.1406
P.1409
P.1410
Terrestrial
Terrestrial
Aeronautical
Terrestrial
24
P.1411
Terrestrial
25
P.1546
Terrestrial
Amount of path loss and delay spread in case of Indoor

propagation
Digital modulation
Land mobile and broadcasting services
High altitude radio communication systems
Coverage and reduction in coverage due to rain in
Broadband radio access
Path loss and delay spread in Out-door short range
propagation
Field strength measurement in Point-to-area propagation
26
27
P.1791
P.1812
Terrestrial
Terrestrial
UWB path loss for indoor/outdoor cases

Field strength calculation in Point-to-area propagation
28
P.1814
Terrestrial
29
30
P.1816
P.2001
Terrestrial
Terrestrial
Calculating beam spreading, absorption and scattering loss

and scintillation in FSO radio communication systems
Urban/suburban planning
Path loss due to signal enhancement and fading
31
P.2040
Terrestrial
Building materials properties and structures
Path loss and diversity enhancement in Digital fixed LOS

links planning
Ionospheric propagation
Estimation of path loss, amount of diversity gain and cross
polarization discrimination (XPD) in Telecommunication
services
Interference calculation
Effects of surrounding environment and path loss
prediction in Broadcasting satellite systems
Fade/fade duration and interference calculation for
Maritime mobile telecommunication
Fade and path loss estimation in Land-mobile
telecommunication
Fade and multipath calculation in Aeronautical mobile
telecommunication systems
253
Frequency of
operation
0.1 GHz-50
GHz
-----------125 MHz15500 MHz
150 MHz- 100
GHz
0.1 GHz- 12
GHz
1 GHz- 55 GHz
--------0.5 GHz- 5.1

GHz
0.8 GHz- 8
GHz
0.8 GHz- 20
GHz
1 GHz- 2 GHz
(sea), 1 GHz- 3
GHz (ground)
30 kHz- 150
kHz
>10 GHz
---------30 MHz-100
MHz
30 MHz-3 GHz
150 kHz-1700
kHz
900 MHz 100 GHz
------------------------------3 GHz- 60 GHz
300MHz- 100
GHz
30 MHz- 3000
MHz
1 GHz- 10 GHz
30 MHz- 3000
MHz
20 THz- 375
THz
0.7 GHz-9 GHz
30 MHz- 50
GHz
----------
Distance
Up to radio
horizon and
beyond
------------0 km- 1800 km
200 km
------All possible
orbital heights
------All possible
orbital heights
All possible
orbital heights
All possible
orbital heights
All possible
orbital heights
0 km- 16000 km
------------100 km- 1000
km
------50 km- 12000
km
Buildings,
interiors
------------------0 km- 5 km
Less than 1 km
1 km- 1000 km
------Up to radio
horizon and
beyond
No limit
------3 km- 1000 km
-------
[20] ITU-R P.619-1(1992),Propagation data required for the evaluation

of interference between stations in space and those on the surface of
the Earth, ITU-R Recommendation
[21] ITU-R P.679-3(2001),Propagation data required for the design of
broadcasting-satellite systems, ITU-R Recommendation
[22] ITU-R P.680-3(1999),Propagation data required for the design of
Earth-space maritime mobile telecommunication systems, ITU-R
Recommendation
[23] ITU-R P.681-3(1997), Propagation data required for the design of
Earth-space maritime mobile telecommunication systems, ITU-R
Recommendation
[24] ITU-R P.840-6(2013),Attenuation due to clouds and fog, ITU-R
Recommendation
[25] ITU-R P.452.15(2013),Prediction procedure for the evaluation of
interference between stations on the surface of the Earth at the
frequencies above about 0.1 GHz, ITU-R Recommendation
[26] ITU-R P.525-2 (1994), Calculation of free-space attenuation, ITUR Recommendation
[27] ITU-R P.526-13(2013),Propagation by diffraction, ITU-R
Recommendation
[28] ITU-R P.530-15(2013), Propagation data and prediction methods
required for the design of terrestrial line-of-sight systems, ITU-R
Recommendation
[29] ITU-R P.684-6(2012),Prediction of field at the frequencies below
about 150 kHz, ITU-R Recommendation
[30] ITU-R P.842-5(2013),Computation of reliability and compatibility
of HF radio systems, ITU-R Recommendation
[31] ITU-R
P.843-1(1997),Communication
by
meteor-burst
propagation, ITU-R Recommendation
[32] ITU-R P.844-1(94),Ionospheric factors affecting frequency sharing
in the VHF and UHF bands (30 MHz-3 GHz), ITU-R
Recommendation
[33] ITU-R P.1147-2(2003),Prediction of sky-wave field strength at
frequencies between about 150 and 1700 kHz, ITU-R
Recommendation
[34] ITU-R P.1321(2013),Propagation factors affecting systems using
digital modulation techniques at LF and MF, ITU-R
Recommendation
[35] ITU-R P.1406-1(2007), Propagation effects relating to terrestrial
land mobile and broadcasting services in the VHF and UHF bands,
ITU-R Recommendation
[36] ITU-R P.1410-5 (2012),Propagation data and prediction methods
required for the design of terrestrial broadband radio access systems
operating in a frequency range from 3 to 60 GHz, ITU-R
Recommendation
[37] ITU-R P.1546-5 (2013) ,Method for point-to-area predictions for
terrestrial services in the frequency range 30 MHz to 3 000 MHz,
[38] ITU-R P.1791-0 (2007), Propagation prediction methods for
assessment of the impact of ultra-wideband devices , ITU-R
Recommendation
[39] ITU-R P.1812-3(2013), A path-specific propagation prediction
method for point-to-area terrestrial services in the VHF and UHF
bands, ITU-R Recommendation.
[40] ITU-R P.1814-0 (2007) , Prediction methods required for the design
of terrestrial free-space optical links, ITU-R Recommendation
[41] ITU-R P.1816-2 (2013), The prediction of the time and the spatial
profile for broadband land mobile services using UHF and SHF
bands, ITU-R Recommendation
[42] ITU-R P. 2001-1 (2013), A general purpose wide-range terrestrial
propagation model in the frequency range 30 MHz to 50 GHz, ITUR Recommendation
[43] ITU-R P.2040 (2013), Effects of building materials and structures on
radiowave propagation above about 100 MHz, ITU-R
Recommendation
[44] ITU-R P.1144-6 (2012), Guide to the application of the propagation
methods of Radiocommunication Study Group 3, Recommendation.
REFERENCES
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
Proceedings
International Telecommunication Union, (ITU), www.itu.int,

2/2/2015, <http://www.itu.int/en/Pages/default.aspx>
Propagation Tutorial, (Mike Willis), 4/2/2015, <http://www.mikewillis.com/Tutorial/propagation .html >
Andrea Goldsmith, Path Loss and Shadowing, in Wireless
Communications, Cambridge, United Kingdom: CUP, 2005.
Radiowave propagation, (ITU 2008), Accessed 2/2/2015,<
http://www.itu.int/rec/R-REC-P/en>
L. K. Bandyopadhyay, S. K. Chaulya, and P. K. Mishra, Mines
Communication Technique, in Wireless Communication in
Underground Mines: RFID-based Sensor Networking, Springer
Science & Business Media, 2009.
Masri, T., Chew, S.P., Wong, C.P., and Lias, K., A study of signal
penetration into building materials, IEEE 1st International
Conference on Computers, Communications, & Signal Processing,
2005. CCSP 2005, Pages: 21-24, Kuala Lumpur, Malaysia.
Sumit Joshi, and Vishal Gupta, A review on empirical data
collection and analysis of Bertonis model at 1.8 GHz, International
Journal of Computer Applications (IJCA), 2012, Volume 56 Number 6, pages: 17-23, October 2012.
Govind Sati, and Sonika Singh, A review on outdoor propagation
models in radio communication, International Journal of Computer
Engineering and Science (IJCES), Volume: 4, Issue: 2, pages: 64-68,
October 2014.
Pooja Rani, Vinit Chauhan, Sudhir Kumar, and Dinesh Sharma, A
review on wireless propagation models, International Journal of
Engineering and Innovative Technology (IJEIT), Volume: 3, Issue:
11, May 2014.
Stylianos Kasampalis, Pavlos I. Lazaridis, Zaharias D. Zaharis,
Aristotelis Bizopoulos, Spyridon Zettas, and John Cosmas,
Comparison of Longley-Rice, ITU-R P.1546 and Hata-Davidson
propagation models for DVB-T coverage prediction, 2014 IEEE
International Symposium on Broadband Multimedia Systems and
Broadcasting (BMSB), pages: 1-4, Beijing, China, June 2014.
Magdy F. Iskander, and Zhengqing Yun, Propagation Prediction
Models for Wireless Communication Systems, IEEE Transactions
on Microwave Theory and Techniques, VOL. 50, NO. 3, pages: 662
673, March 2002.
Tapan K.Sarkar, Zhong Ji, Kyungjung Kim, Abdellatif Medouri, and
Magdalena Salazar-Palma, A survey of various propagation models
for mobile communication, IEEE Antennas and Propagation
Magazine, Volume: 45, Issue: 03, pages: 239 307, June 2003.
ITU-R P.1238-5 (2007), Propagation data and prediction methods
for the planning of indoor radio communication system and radio
local area networks in the frequency range 900 MHz to 100 GHz,
ITU-R P.1411-7 (2013) ,Propagation data and prediction methods
for the planning of short-range outdoor Radiocommunication systems
and radio local area networks in the frequency range 300 MHz to 100
GHz, ITU-R Recommendation
ITU-R P.528-3 (2012),Propagation curves for aeronautical mobile
and radio navigation services using the VHF, UHF and SHF bands,
ITU-R P.682-3 (2012), Propagation data required for the design of
Earth-space aeronautical mobile telecommunication systems , ITUR Recommendation
ITU-R P.1409-1(2012),Propagation data and prediction methods for
systems using high altitude platform stations and other elevated
stations in the stratosphere at frequencies greater than about 1 GHz,
ITU-R P.618-11(2013),Propagation data and prediction methods
required for the design of Earth-space telecommunication systems,
ITU-R P.531-12(2013),Ionospheric propagation data and prediction
methods required for the design of satellite services and systems,
254
Proceedings
Issues of Alpha Mechanism in Subsonic Wind Tunnel

Testing for a Hybrid Buoyant Aircraft
Experimental Fluid Dynamics
A.U.Haque*, W. Asrar, E. Sulaeman, J.S. Mohamed Ali and A.B. Embong

IIUM-LSWT, International Islamic University Malaysia (IIUM)
Kuala Lumpur, Malaysia
*
anwar.haque@live.iium.edu.my
Abstract In subsonic wind tunnels with large test sections,

scaled down models are attached to the balance with the help of
model support system. Such a system mainly includes a main
strut and an auxiliary strut, also known as pitch rod. Flow
measurements of such a system are required to be subtracted
from the raw data of the wind tunnel. For this purpose, it is
important to find the contribution of model support system
separately. Potential issues and limitations related to such testing
are discussed in the light of technological constraints. Special
focus is given on the streamlined profile of main strut and wind
tunnel testing of a hybrid buoyant aircraft. In this paper we
experimentally simulate a situation in which height of both these
struts are kept constant. Wind tunnels tests are conducted to find
the aerodynamic and stability characteristics at constant angle of
attack as well as yaw angle. An asymmetric pattern is found in
the pitch and yaw stability for a defined array of beta sweep.
Such a support system not contributes towards drag but also in
lift as well as in side force.
Index Terms Wind Tunnel Testing, Pitch Rod,
Aerodynamic and stability measurements, Data acquisition and
reduction system, Strut Arrangement
I. INTRODUCTION
It is well known that the experimental data obtained from a
wind tunnel is subject to different corrections. Correction due
to model support group is among one of important correction
for which wind tunnel is run for model off and strut on
conditions. Struts are usually geometrically tapered with less
diameter and chord at the tip than that at the root which is
attached with the balance. Moreover, they are of cylindrical or
streamlined shapes. Such struts have more frontal area and
have complex interaction of the flow when scaled down model
is attached at the tip of it. One of the prospective solutions is to
have tapered struts of profile similar to a diamond shaped
airfoil. Diamond shaped struts are usually used in supersonic
[1-3] and hypersonic wind tunnels [4-5] in which there is
requirement to hold the model from the outer profile and no
interaction of the strut with the base of the model. However, to
the authors best knowledge, its utilization in subsonic speed
has hardly been seen in the literature. Most of the earlier work
done is to compare the effects of different arrangement of
streamlines, round or cylindrical struts with pitch rod at fixed
position of angle of attack [6-7]. But such a method is not
255
quick measuring method as the individual has to change the

position of pitch rod at every required angle of attack and
hence forming different V which require additional fillets to
be installed for all angular setting of the struts tested
individually [6]. Also, in comparison with the main strut the
dimensions especially the diameter and the thickness of the
pitch rod is usually quite less. One end of the pitch rod is
usually connected with the gear system of alpha mechanism
which allows the rod to swing in a plane perpendicular to the
turntable axis. If the tip of the pitch rod exposed to the free
stream velocity is not physically attached with the main strut
than there might be some effect of flow induced vibration on
the overall results.
The drag of a model mounted only on main strut in a wind
tunnel is always more than the individual sum of the isolated
strut and isolated body[8]. Isolated strut are used when model
is either fixed at certain position or the operator manually
change the position with the help of instruments like
inclinometer. Also it is usually used when the geometric
dimensions of the scaled down model quite small to have
auxuilary strut to be position at the aft part of the model.
Examples of few models tested earlier on single strut are
shown in Fig. 1.
It is important to take separate measurement of the strut
arrangement to get a more general picture of the estimated
parameters of main strut with axillary strut for defined range of
side-slip angle and angle of attack. Such a strut arrangement
can cause complex interference flow fields including the
increase in drag and pitching moment. This work aims to
quantify the magnitude of such effects and to give a first and
knowledge of general arrangement of the experimental setup
for fixed height of the pitch rod.
Due to more interference and drag due to strut arranged in
tandem [9] form and further discussion on it or other two or
three strut arrangement is beyond the scope of present work.
In early times, the manufacturing and assembly techniques
for wind tunnel models are not so sophisticated as that of the
present time. Moreover, the size of the test section is also
small. Such technological constrains limits the overall length of
the model and pitch rods has to positioned at some inclined
angle to attached its tip with the base of the model. In order to
define any array of input angle of attack; the pitch rod has to
move in a plane perpendicular to the tunnels floor.
Proceedings
auxiliary strut increases for negative angle of attack and

decreases for positive angle of attack. In this way there will be
multiple numbers of cases to estimate the contribution of
model support only. One of the optimum solutions to keep
minimum number of blow downs is to keep the height of the
pitch rod equal to that of the main strut. Complete details of
this prospective solution are discussed below with the help of
an example of on-going wind tunnel testing program of a
hybrid buoyant aircraft. Such aircraft are powered by fuel and
has the ability to generate certain amount of aerodynamic lift
for the weight which is not balanced by the aerostatic lift.
II. PROSPECTIVE SOLUTION
In the case of aircraft the operational angle of attack is
usually limited to 16-18 degree on the positive side and -6
degree on the negative side. Hence the cases for experimental
investigation for aerodynamic and stability characteristics
through wind tunnel testing will be more for the positive one.
In order to define a positive angle of attack, the alpha
mechanism pushes the pitch rod in the downward direction. By
doing so, the tail of the aircraft moves towards the tunnel floor
and it is vice versa for the situation where negative angle of
attack is required. In comparison with conventional aircraft,
hybrid buoyant aircraft may need to fly at altitude less than that
of pressure height [10]. Although a lifting fuselage can provide
additional lift and hence high lift to drag ratio for the cruise and
take-off segment of the flight but there may exist a scenario in
which a pilot may need to give negative angle of attack [11]. In
order to simulate such conditions in wind tunnel, there is
always a requirement to do more testing at negative angle of
attack for which the pitch rod has to move up to attain a height
equal to that of the main strut. For data reduction; all the
aerodynamic and stability parameters has to nondimensionalized by reference dimensions of the wing. All such
dimensions, obtained from the CAD model (actual as well as
scaled down) are tabulated below in Tab. 1 for quick reference.
A schematic view of the model attached with the diamond
shaped strut and pitch rod attached at its base is shown in Fig.
2. It is important to note that the reference area value is
obtained through an iterative analysis work for which lift from
the fuselage is added into the main wing by using an equivalent
wing with an airfoil as that of the wing [12].
(a) Delta wing model
(b)
Airship Model
TABLE. 1 REFERENCE DIMENSIONS OF HB AIRCRAFT

Wing
Area
2
(c)
Span
Chord
Origional
27.5
20
1.5
Scaled down
0.086
0.075
Hybrid lifting hull model
Fig. 1. Few models tested in IIUM-LSWT
Now days, through precise machining, scaled down models

in large dimensions can be manufactured by using combination
of metal and composite. For such models, the height of
256
It is pertinent to highlight that the scope of present research

work of model support system is not only limited to the
aerospace application like that of hybrid buoyant aircraft,
discussed above. But can also be applied to any body in which
there is requirement to evaluate the effect of defining the body
at some angle to the free stream velocity. Estimation of
parachutes drag [13], birds aerodynamics [8,14], cars

aerodynamics [15-16], study of lift generated wakes [17] are
few examples of it.
Proceedings
F is an aerodynamic force) can be expressed in the nondimensionalized form, Eq. (1). In this relationship, instead of
taking the actual exposed area of the body, the platform area of
the wing is usually used as S.
CF
F
1 V 2 S
2
(1)
Where, CF F,V, S and is coefficient of aerodynamic force,

aerodynamic force, free stream velocity, reference area and
density of air respectively. Similarly, the aerodynamic
moments also can be expressed in the form of non-dimensional
coefficients. Since a moment is the product of a force and a
length it follows that a non-dimensional form for a moment is
CQ
Fig. 2. Isometric view of HB aircraft installed in test section
Q
1 V 2 Sl
2
(2)
Where, Q is any aerodynamic moment about center of gravity

and l is a reference length.
III. EXPERIMENTAL SETUP
In order to find the aerodynamic and static stability
behavior of model support system discussed in the previous
section, tests are conducted in IIUM-LSWT at 40 m/s. IIUMLSWT is a closed-loop wind tunnel with a test section of
dimensions 1.5m 2.3m 6m and maximum speed of 50 m/s
[18]. A pictorial view of the test section is shown in Fig. 4(a)
[19]. In this tunnel, the balance section is placed on the floor to
have some height for the external balance. Both the struts
(main as well as auxiliary i.e. pitch rod for alpha mechanism)
are attached with the balance. The same can be observed from
Fig. 4(b) in yellow colour. This wind tunnel is currently one of
the largest wind tunnels in Malaysia and has excellent flow.
The limitations of this experimental setup and its
repurcation of the data obtained from DARCS system of the
tunnel are as follows:
a. It is mandatory to have the pitch rod for defining angle of
attack range. Consequently, in order to quantify their
aerodynamic and stability effects; no comparisons are made
between results from diamond shaped strut with and
without the pitch rod strut.
b. No tests are performed with the model attached to the
defined model support system. Therefore the investigation
reported in this paper is limited to the estimation of struts
along contribution and interference arising from
combination of model-strut arrangement is beyond the
scope of presented work.
c. No tests for angle of attack are possible as the model
support system is rigidity fixed with the balance.
(a) Side view
IV. EXPERIMENTAL SETUP
(b) Front View

Fig. 3. Positive incidence of the model with the help of pitch rod
For such cases, it is simply replacing the terms of reference

area for aerodynamic forces and reference dimension for the
moments. For example the non-dimensional quantity F (where
257
In order to evaluate the effectiveness of the proposed

engineering method, Fig. (5); wind tunnel blow downs are
carried out at different velocities. All the results are plotted in
the form of graph to get first-hand knowledge about the general
trends of aerodynamic and stability characteristics of model
support system. First the wind tunnel is run for a series of blow
downs (=0, =0) by increasing V from 30 to 50 m/s and with

step velocity of 5 m/s.
Proceedings
almost constant. However, an increase in CL at V=45 m/s and

increase in CD at V=40 m/s are observed.
0.14
CL
0.12
CD
0.1
Cm
Cy
Coefficients
0.08
Cn
0.06
0.04
0.02
0
30
35
40
45
50
-0.02
-0.04
V (m/s)
Fig. 6 Variation in aerodynamic and stability coefficients with increase in the

free stream velocity
(a)
IIUM-LSWT
50
Lift
40
Forces (N)
Side
Drag
30
20
10
0
-20
-15
-10
-5
10
15
20
(degree)
-10
-20
-30
(b)
-40
Balance Compartment
(a) Overall aerodynamic forces

Fig. 4 Details of IIUM-LSWT
0.6
CL
0.5
Coefficients
Cn
CD
0.4
0.3
0.2
0.1
0
-20
-15
-10
-5
0
-0.1
10
15
20
(degree)
-0.2
-0.3
-0.4
-0.5
(b) Overall aerodynamic coefficients

Fig. 7 Aerodynamic properties of model support system
Fig. 5 Proposed arrangement of model support struts
All the moments and its coefficients are estimated about the
moment reference centre (MRC) of the balance. Values of
It can be observed from Fig. 6 that for the said range of

velocity, all the aerodynamic and stability coefficients remain
258
increases till for = 5 but a drastic decrease in its value is

observed till = 150, Fig. 8(b). But its magnitude is not
consistent with that of its value at = 150. Values of Cn for .
200 are almost constant but a sudden increase in its value is
found at = 50. Since the overall behavior of these two
stability coefficients is non-linear. Therefore, it is quite
difficult to comment on the overall slopes of C m and C n .
115
Moment (N.m)
75
55
35
15
(a)
-10
-5
-5 0
(degree)
10
15
Plots of pitch and yaw moment due to sweep
Cm
0.8
0.6
0.4
0.2
0
-10
-5
10
15
20
(degree)
(b)
-15
-10
-5
-0.005
10
15
20
-0.01
-0.015
-0.02
(degree)
(c) Cn vs
Fig. 8 Stability characteristics of model support system
Sometimes, the aerodynamic force of the body alone is defined

as the difference between the aerodynamic force of the body
attached with the strut and results of strut alone testing.
However, this is not the true representation as there is always
some effect of the interference due to model support system as
well. For example, the complete analytical relationship of the
drag force is expressed as follows [8]:
(1)
For measurements of aerodynamic and stability coefficients

of only model support system placed normal to the flow, a
method is used where both the struts are at same height and
attached through a thin wire. For defined range of velocity, all
the aerodynamic and stability coefficients remain almost
constant for = 00. Diamond shaped profile of the mains strut
acts as a symmetric airfoil, which result in significant
contribution towards the side force. For side force rest of the
forces and moments, the overall trend is not symmetric.
Specially, an oscillating behaviour is observed in the trend of
Cn vs . Such effects of the measured aerodynamic and
stability characteristics caused by main strut and the pitch rod
are of considerable magnitude and should be excluded from the
test results of complete configuration. Although the presented
results of proposed experimental setup furnish some interesting
finding but its interference with the model deserved much more
attention.
20
-15
0
-20
V. CONCLUSION
1.2
-20
0.01
0.005
Where, Dm, Ds and DI are drag of the body, strut and due to
interference effects respectively. Drag due to tare is usually
automatically subtracted in online calculations by the DARCS
and hence it is not account far in the above equation. DI would
take out any interference effects due to supports and can be
estimated in future by subtracting the model-on wind-on
condition from the model off-wind on condition. This is a
shortcut (not perfect) to avoid lengthy procedure involving
measurements with inverted model [20].
Yaw Moment
-15
0.02
0.015
Dm DB DS DI
Pitch Moment
95
-20
0.025
Cn
coefficient of yaw and side force are negative, whereas the sign
of pitching moment is positive. For sweep test; the model
strut assembly remained rested on tunnel turntable and rotated
at -20o to +20o range with 50 increments. As more attention is
given for defining the problem description than the flow
conditions. Hence, this technical note does not attempt to
predict the aerodynamic and stability coefficients for beta
sweep at different free stream velocities. A drastic increase in
side force is observed at = 150, Fig. 7(b). This figure also
shows a continuous increase in CL for defined range of and a
slight droppage in value of CD for = 100. However, after =
100, CD increases linearly till = 200 .
An asymmetric pattern is observed in overall trends of pitch
and yaw moment, Fig. 8(a). Same thing is reflected in
corresponding coefficient plots as well i.e. Cm vs and Cn vs
plots, respectively. However, general behavior is more
sinusoidal for the plot of Cn vs . Fig. 8(c). Value of Cm first
Proceedings
Cm vs
259
ACKNOWLEDGMENT
[13]
The support of the Ministry of Science, Technology and

Innovation (MOSTI), Malaysia, under the grant 06-01-08SF0189 is gratefully acknowledged. Authors are thankful to
Dr. Fadhil Hasim and his team for providing the diamond
shaped strut for the experimental work.
[14]
REFERENCES
[15]
[1]
[2]
[3]
[4]
[5]
[6]
[7]
D., Biermann, & W. H. Herrnstein, The interference between

struts in various combinations. NASA Report No. NASA-L468, 1933
E. P., Jones, S. M., Townsend, Y., Guy, & T. E. McLaughlin,
Reduction of the wave drag of a blunt body by means of a
standoff spike. AIAA Aerosp. Sei. Meet., 38th, Reno, NV,
2000
S. Feldman, Some shock tube experiments on the chemical
kinetics of air at high temperatures. Journal of Fluid
Mechanics, 3(03), 225-242, 1957.
K. L., Mikkelsen, & D. F. Long, Refurbishment, Calibration,
and Initial Testing in the ASE Hypersonic Wind Tunnel. In
43rd AIAA Aerospace Sciences Meeting and Exhibit, 2005.
[16]
[17]
[18]
M. H. Bertram, (1950). Investigation of the pressure-ratio

requirements of the Langley 11-inch hypersonic tunnel with a
variable-geometry diffuser, 1950.
D., Biermann, & W. H. Herrnstein, The interference between
struts in various combinations. NASA Report No. NASA-L468, 1933
V., Nguyen, Y., Drolet, & Watt, G. (1995, January).
Interference of various support strut configurations in wind
tunnel tests on a model submarine. In 33rd Aerospace
Sciences Meeting and Exhibit (p. 443).
[8]
V.A, Tucker, "Measuring aerodynamic interference drag

between a bird body and the mounting strut of a drag
balance." The Journal of Experimental Biology 154.1 (1990):
439-461.
[9]
R. C. Pankhurst, (1958). Wind-Tunnel Interference Due to

Model Supports. Aircraft Engineering and Aerospace
Technology, 30(2), 55-55.
[10]
A.U., Haque, W. Asrar, A.A, Omar, E., Sulaeman and J.S,

Ali, "Assessment of Engines Power Budget for Hydrogen
Powered Hybrid Buoyant Aircraft", Journal of Propulsion
and Power Research (in press), 2015
[11]

Ali, "Aerostatic and Aerodynamic Modules of a Hybrid
Buoyant Aircraft: an Analytical Approach", IIUM
Engineering Journal, IIUM Engineering Journal, Vol 16, No.
1 (2015).
[12]

Ali, Power-off static stability analysis of a clean
configuration of a hybrid buoyant aircraft, 7th Ankara
Internatonal aerospace conference, 11-13 September 2015 METU, Ankara Turkey.
260
Proceedings
J. A., Weiberg, & K. W. Mort, Wind-tunnel tests of a series

of parachutes designed for controllable gliding flight.
National Aeronautics and Space Administration, 1967
G. R.,Spedding, M.Rosn, & A. Hedenstrm, A family of
vortex wakes generated by a thrush nightingale in free flight
in a wind tunnel over its entire natural range of flight speeds.
Journal of Experimental Biology, 206(14), 2313-2344, 2013
A. Cogotti, Ground effect simulation for full-scale cars in the
pininfarina wind tunnel (No. 950996). SAE Technical
Paper.1995
B. Hetherington, and D. B. Sims-Williams, Support Strut
Interference Effects on Passenger and Racing Car Wind
Tunnel Models. No. 2006-01-0565. SAE Technical Paper,
2006.
J., Rossow, J. N., Sacco, P. A., Askins, L. S., Bisbee and S.
M Smith, Wind-tunnel measurements of hazard posed by liftgenerated wakes. Journal of aircraft, 32(2), 278-284, 1995.
F. Hasim, R. Rusyadi, W. I. Surya, W., Asrar, A. A.Omar, ,
J., Syed Mohamed Ali, & R. Kafafy, The IIUM low speed
wind tunnel. Proceedings of EnCon2008 2nd Engineering
Conference on Sustainable Engineering Infrastructures
Development & Management, December 18-19, 2008,
Kuching, Sarawak, Malaysia
[19]
Users Manual 6-Component External Balance for

International Islamic University Malaysia Wind Tunnel, S &
V Teknik Sdn. Bhd and Aero Engineering, 2015.
[20]
J. Barlow, W. Rae, A. Pope, Low Speed Wind Tunnel

Testing, Jon Wiley & Sons, 1999.
Proceedings
Fairing Separation Dynamic Analysis Using

Analytical Approach
Muhammad Tanveer Iqbal
Dr. Abdul Majid

Islamabad, Pakistan
mrtanveeriqbal@gmail.com

Islamabad, Pakistan
majid363@yahoo.com
Abstract Fairing is the upper portion of launch vehicle,

usually made in two halves. It is used to protect the satellite from
loadings (thermal, aerodynamic and acoustic) of harsh
environment during ascent. Separation of fairing at particular
altitude is most critical part of launch vehicle flight which
dictates the success or failure of the whole mission.
time. The high spring stiffness will produce the high angular
velocity in fairing halves and cause its collision with launch
vehicle. Therefore, optimized value of stiffness must be
evaluated.
This paper presents the designing of spring for separation

system of fairing using 1dof dynamic equations. Material
properties of base ring and spring are considered and integrated
in dynamic system of equations. These equations are modeled
and evaluated. Effects on separation time due to variations in
spring stiffness and compressed length are analyzed. After
analysis , best spring parameters are selected on the basis of
fairing separation requirements. Developed analytical method is
also validated through professional fairing separation software.
Keywordssatellite launch vehicle; fairing jettison; heat
shield; dynamic analysis; analytical approach; separation springs
I.
INTRODUCTION
Fairing is designed to shield the satellite from acoustic,

aerodynamic and especially thermal loads. To minimize these
loadings inside the fairing, honeycomb and carbon fiber
sandwich structure is frequently employed. Separation of
fairing is typically done at altitude where free molecular heat
flux is less than 1135W/m2 to avoid damages to satellite [1-3].
Bending load and dynamic pressure are not considered for
fairing separation as their utmost values are achieved prior in
the flight [4].
In this paper, along with stiffness, other design parameters

of spring used in fairing jettison mechanism are also evaluated.
For this purpose, analytical equations for rotational dynamics
are derived using laws of motion. Then using Range-Kutta
method, these equations are solved, simulated and analyzed.
Acquired results are validated through professional software
used for fairing separation calculations..
II.
FAIRING SEPARATION MECHANISM
Jettison is started by lateral separation mechanism that

unlocks the fairing bottom from launch vehicle. Then pyrobolts of longitudinal separation mechanism are exploded that
separate both halves of fairing, followed force is released by
compressed springs which rotate the halves about hinges. After
reaching certain angle, fairing starts free fall rotation about its
center of gravity. Fairing and hinge structures under
consideration are illustrated in Fig. 1(a) and 1(b).
Successful mission is heavily depends on fairing separation.

Many missions were failed just due to no or late separation of
fairing. In the March 4, 2011, Glory climate satellite launched
by Taurus launch vehicle was failed to achieve specific orbital
altitude due to failure of fairing separation and collapsed back
to earth [5].
Two halves of fairing are separated by using pyro-bolts and
gas/spring thrusters. In the given case, spring thrusters are
selected for designing jettison mechanism. Two main
considerations for designing separation mechanism are that the
magnitude of force produced by spring system must be enough
to separate the fairing's halves in given time at very high
acceleration and after separation, there should be no collision
of opened halves with the launch vehicle [6]. Both are directly
related to spring stiffness. If spring stiffness is low, sufficient
force would not be produced to open the fairing halves in given
261
Fig. 1(a): Fairing structure
Proceedings
For left half of fairing, summation of all the moments about

the hinge point is given by:
(1)
After re-arranging (1);
(2)
Fig. 1(b): Fairing model along with spring mechanism and hinge structure
III.
FAIRING SEPARATION DYNAMICS
Dynamic equations are derived for 1-DOF system. For

analytical derivation, fairing is considered as rigid body.
During separation, hinge point's linear movement is neglected.
Frame moves with constant linear acceleration fixed on launch
vehicle, is taken as reference [7]. Pitch angle of launch vehicle
is taken as zero. Free body diagram of left half of fairing before
fairing separation is illustrated in Fig. 2. Free body diagram of
left half of fairing during separation is shown in Fig. 3.
Where, 'Izz ' is the moment of inertia of left half about

hinge. 'Fs ' is the magnitude of force applied by all springs. 'ds'
is the moment arm of spring force 'Fs '. 'Mfair' is the mass of
fairing half. 'a' is the linear acceleration applied on the fairing
due to instantaneous thrust of launch vehicle, which is
considered as constant during separation. ' ' is the angle of
rotation of fairing half. 'Fs' is calculated using equation given
below.
(3)
Where, 'K' is stiffness of one spring. 'Xtot' is the total
compressed length of a spring. 'Ns' is the number of springs. In
given case, 2 springs are considered in separation mechanism.
'Xdef' is the instantaneous uncompressed length of spring during
separation and it is found using cosine law of triangle given by,
(4)
Equations (1) to (4) are coupled through ' '. These interlinked dynamic equations are solved by using Range-Kutta
method. Same is used for dynamic solution of right half of
fairing with trivial modifications.
Theoretically, force exerted by the spring system 'Fs ' on the
base ring and the reaction force 'Fr ' generated by the ring on
the spring system is equal. But in practical, "action is not equal
to reaction" due to inelastic contact of two bodies [8-9]. Spring
impingement on the base ring causes the absorption of energy
in the ring. Amount of energy absorbed depend on the
materials of spring and ring. Thus reaction force produced by
ring will always be less than the spring force. Harder material
absorbs less energy and vice versa. Force loss factor 'e' is
considered to make analytical solution closer to real results.
Therefore 'Fr ' is used instead of 'Fs ' in equation (2).
Fig. 3: Free body diagram of left half of fairing before separation
(5)
Equations (1) and (2) are valid prior to fairing separation
from hinge. After fairing separation, analogous equations are
required to derive for motion of fairing about its center of
gravity.
IV.
Equations are implemented and solved in MATLAB.

Spring parameters including spring stiffness, spring material
and compressed length, are evaluated to satisfy given
separation requirements given in table 1.
Fig. 4: Free body diagram of left half of fairing during separation
262
Proceedings
TABLE 5: EFFECTS ON SEPARATION TIME DUE TO VARIATION IN SPRING

STIFFNESS
TABLE 1: FAIRING SEPARATION REQUIREMENTS
Fairing Jettison Angle
70 deg
Fairing Jettison Time
< 0.8 sec
Typical values of system parameters used in developed

methodology are given in table 2.
S.
No
Spring
Stiffness
(N/m)
Spring
Compressed
Length (m)
Max. Reaction Force

'Fr' on each half of
Fairing (N)
Separation
Time (sec)
70,000
0.08
11066
0.96
80,000
0.08
12646
0.88
90,000
0.08
14228
0.83
100,000
0.08
15808
0.78
TABLE 2: SYSTEM PARAMETERS
On the basis of results, spring having stiffness of 100,000

N/m and compressed length of 0.08m is selected since it
fulfills fairing separation requirements with sufficient margin.
Fig. 5 shows the effect of stiffness and compressed length of
spring on separation time of fairing.
Mass of Fairing Half
137 kg
Moment of Inertia of one half about hinge
700 Nm2
Location of Xcg
1.75 m
Location of Ycg
0.39 m
Linear Acceleration of Fairing due to Thrust
46.44 m/sec2
Gravitational force at Separation Altitude
9.464 m/sec2
Moment Arm of Spring
0.61 m
Earlier discussed spring loss factor is evaluated from

professional mechanics software. Values for different
materials of spring and base ring in-contact are given in table
3.
TABLE 3: SPRING FORCE LOSS FACTOR
Material of Spring
Material of Base Ring
Force Loss Factor
Aluminum
Aluminum
0.09
Aluminum
Steel
0.012
In our case, 0.012 is used because material considered for

spring is aluminum and for base ring is steel. Different cases
are simulated and analyzed for various spring constants and
spring compressed lengths given in table 4 and 5.
Fig. 2: Effects of Spring Parameters on Separation Time
For validation, results computed by analytical method are

compared with professional fairing separation software. Fig. 6
and 7 authenticates analytical method.
TABLE 4: EFFECTS ON SEPARATION TIME DUE TO VARIATION IN SPRING

COMPRESSED LENGTH
Spring
Compressed
Length (m)
Max. Reaction Force

'Fr' on each half of
Fairing (N)
0.04
7904
0.05
9880
S.
No
Spring
Stiffness
(N/m)
100,000
Separation
Time (sec)
No
Separated
No
Separated
0.06
11856
1.1
0.07
13832
0.89
0.08
15808
0.78
Fairing's half
separated from
hinge
Fig. 3: Angular Velocity about Z-axis of Left Half of Fairing
263
Proceedings
Path of CG
Fig. 4: Reaction Force of Left Half of Fairing
Fig. 6: Fairing's half Movement during Separation
is presented in Fig. 8. At 0.78

Fairing's opening angle
sec, fairing achieved rotation of 70o so it is released from
hinge, satisfying the requirements given in table 1.
V. CONCLUSION
Results show that designed analytical approach for fairing
separation dynamic analysis is fairly close to professional
software approach. Inclusion of material properties of spring
and base ring made results more comparable. Among various
springs of different stiffness and compressed lengths, best one
(having stiffness of 100,000 N/m and compressed length of
0.08m) is chosen on the basis of given fairing separation
parameters. 3dof and 6dof systems will be considered; modeled
and simulated in future that will provide complete dynamics of
fairing separation along all axes.
Fairing's half
separated from
hinge
ACKNOWLEDGMENT
We would also like to show our gratitude to Dr. Imran
Afzal for providing insight and expertise that greatly assisted
the research.
REFERENCES
[1]
Fig. 5: Opening Angle Theta of Left Half of Fairing
[2]
[3]
Fig. 9 presents the movement of fairing half in body

frame. Red curve shows the cg variation of fairing's half
during separation.
[4]
[5]
[6]
[7]
[8]
[9]
264
Steven J. Isakowitz, Joshua B. Hopkins, Joseph P. Hopkins Jr.,

International Reference Guide to Space Launch Systems, 4th ed.,
AIAA, pp. 341, 2004.
Soyuz User's Manual, Ariane Space, vol 2 rev. 0 ,March 2012.
Atlas Launch System Mission Planner's Guide, Atlas V
Addendum(AVMPG), International Launch Services, pp. 2-7, Dec.
1999.
Antonio Pagano, Global Launcher Trajectory Optimization for Lunar
Base Settlement, Delft University of Technology, May 2010.
spacenews.com,Orbitals-glory-launch-failure-review-nears-conclusion,
2011. [Online]. Available: http://spacenews.com/orbitals-glory-launchfailure-review-nears-conclusion/. [Accessed: Aug. 9, 2015].
Choong-Seok Oh, Byung-Chan Sun, Yong-Kyu Park and Woong-Rae
Roh, Payload Fairing Separation Analysis Using Constraint Force
Equation, International Conference on Control, Automation and
Systems, Gyeonggi-do, Korea, Oct. 27-30, 2010.
Raymond H. Schuett, Bruce A. Appleby, Jack D. Martin, Dynamic
Load Analysis of Space Vehicle System ,Launch and Exit Phase,
General Dynamics Convair Division, Report No. GDC-DDE66-012,
June 1966.
W. Bauer, G.D. Westfall, Physics for Scientists and Engineers, 3rd

ed., McGraw Hill, 2008.
en.wikipedia.org, 'Inelastic Collision', 2015. [Online]. Available:
https://en.wikipedia.org/wiki/Inelastic_Collision. [Accessed: Aug. 8,
2015].
Proceedings
Launch Vehicle Control based on Dynamic

Inversion with Sliding Mode Neural Network
augmentation
Saqib Alam
Syed Minhaj un Nabi Jafri
SUPARCO, Pakistan
e-mail: alam.saqib@gmail.com
SUPARCO, Pakistan
e-mail: jafri_minhaj@yahoo.com
xd = Dist. from nose to vehicle center of pressure
zr = Dist. from central line of vehicle body to parallel
central line of thrust pipe
xz = Dist. from nose to vehicle center of gravity
V = Air stream velocity of vehicle
C = Side force
L= Lift force
D = Drag force
= Pitch angle
= Yaw angle
= Roll angle
= Flight Path Angle
= Local flight path angle
= Heading angle
Abstract In this paper, a robust attitude control scheme for

satellite launch vehicle (SLV) is presented that is based on inverse
dynamics control augmented by sliding mode neural network
compensator. In this proposed scheme, the inverse dynamics
control acts as the primary control agent that computes the
control moment required to not only stabilize the launch vehicle
but also to implement the steering commands obtained from
guidance scheme. This control action is derived solely by the
inversion of launch vehicle dynamics; therefore, it is vulnerable to
errors resulting from the uncertainties, simplifying assumptions,
nonlinearities and various local and environmental disturbances
all of which cannot be accurately modeled. This problem is
addressed by the secondary control agent of the proposed control
scheme i.e. a sliding mode neural network compensator. This
nonlinear compensator is independent of the system dynamics and
therefore ensures stability against the un-modeled dynamics and
the disturbances acting on the launch vehicle. Along with this
hybrid control architecture, a guidance scheme is also developed
that reshapes the reference attitude profile on the basis of any
deviation from the desired trajectory of the launch vehicle so that
the desired orbit is reached. Lyapunov stability criterion is
incorporated in the control loop to guarantee system stability and
the efficacy of guidance and control scheme is verified on a six
degree of freedom simulation model developed exclusively in
Matlab/Simulink using reference data for a four stage launch
vehicle. In these simulations, the launch vehicle is subjected to
various severe combinations of disturbances and parametric
variations and the control system successfully compensates all the
disturbances and efficiently tracks the desired attitude profile and
attains the targeted orbital parameters.
KeywordsLaunch Vehicle Control, Dynamic Inversion,
Inverse Dynamic Control, Guidance and Control, Sliding Mode
Neural Network Compensator
NOMENCLATURE
m = Mass of vehicle
= Angle-of-Attack
= Side slip angle
= Deflection angle
J = Moment of inertia
= Angular rate
g = Gravitational acceleration
P = Thrust force
xr = Dist. from nose to nozzle center of gravity
I. INTRODUCTION
The access to space in a reliable and cost effective way is a
critical requirement these days. This ability depends largely on
the performance of the guidance and control system of the
launch vehicle. The classical controllers for the ascent phase of
SLV depend heavily on the accuracy of its dynamic modeling
in order to perform as anticipated. Because, it may not satisfy
the stability and performance requirements if the SLV is
outside of its desired trajectory envelop. But, due to the various
uncertainties, nonlinearities and disturbances acting on the
system, the dynamics of SLV cannot be precisely modeled.
Therefore, the controller needs to have the ability to perform
even with inaccurate dynamic model. The controller of a
launch vehicle has to satisfy three often contradictory
requirements that are to stabilize the vehicle, reduce trajectory
deviations by efficiently implementing the steering commands
from guidance scheme and to minimize the angle of attack in
the high dynamic pressure region to avoid structural overload.
During the last few years, a lot of work has been reported in
the field of adaptive and nonlinear control of launch vehicle [1]
[2] [3] [4] [5]. Such techniques have the adaptability to adjust
their output in accordance with the varying dynamics of the
system while retaining the desired performance and stability.
Similarly, various techniques have been devised to enhance the
robustness of the inverse dynamic controller [6] [7] against
modeling uncertainties.
265
This research is aimed at demonstration of the simplicity

and effectiveness of inverse dynamic controller when it is aided
with a sliding mode neural network compensator in developing
the autopilot of SLV. The advantage of using this architecture
is that the primary control produces just enough control
moment that is required for vehicle stabilization and trajectory
following resulting in a simplified and efficient control law.
Whereas all the additional factors are taken care of by the
secondary control, hence ensuring robustness and adaptability.
The weights of the sliding mode neural network are tuned
adaptively in synchronization with the changing environment.
In order to inject a satellite in a desired orbit, an external
guidance loop is also required alongside the inner control loop
to provide the steering commands to the controller on the basis
of trajectory deviations that may occur due to environmental
and local disturbances. A simplified guidance scheme is
proposed here for successful accomplishment of orbital
parameters.
To analyze ascent flight trajectory, 6DOF simulation model

is developed in Simulink MATLAB in which the data of four
stage SLV is incorporated. The symbols definition of SLV is
depicted in Fig.1.
Mass Flow rate

(kg/s)
265
103
37
13
Burn Time (s)
70.7
53.8
46.9
46.9
Stage Dia. (m)
1.3
1.0
0.7
0.5
The launch vehicle experiences various aerodynamic forces

and moments during its flight that act as disturbances and cause
trajectory deviations. In order to simulate these effects,
trajectory of 500 km orbit is considered with injection velocity
of 7612 m/s2, DATCOM is used to compute aerodynamic
coefficients, mass flow rate and thrust are considered as
constant. The equations of motion are formulated considering
the rigid body dynamics, and are given as follows [2]:
mV mg sin cos 4P cos cos D
mV mg cos 4 P sin cos
The paper is organized as follows. Section II illustrates the

mathematical modeling of SLV. Section III describes the
design of inverse dynamic control with sliding mode neural
network compensator for attitude control of SLV along with
the guidance loop. The simulation results are presented in
Section IV and Section V provides the concluding remarks.
II. DYNAMIC MODELING OF SLV
Proceedings
2 P cos L
(2)
mV mg sin sin 4 P sin

2 P cos C
x
y
( J J z ) yz
m qS l 2
Pzr
dx m k y
Jx
VJ x
Jx
m qS l 2
(Jz Jx )
C ( xd xz )
xz
dy m k
Jy
Jy
VJ y
2 P ( xR x z )
Jy
(Jx J y )
Jz
y x
L( xd xz )
m qS l 2
dz m k
Jz
VJ z
2 P ( xR x z )
Jz
(3)
(4)
(5)
(6)
Where,
Fig. 1 Symbols definitions
Table I depicts the four stage data used to simulate the

trajectory of SLV.[1]
TABLE I.
(1)
LAUNCH VEHICLE DATA
Parameters
1st stage
2nd stage
3rd stage
4th stage
Liftoff Mass (ton)
31.34
9.31
2.87
0.91
Fuel Mass (ton)
18.76
5.55
1.72
0.54
Thrust (kN)
593
262
94
30

III. DESIGN OF IDSNN CONTROL
For Launch vehicle, the proposed control architecture
266
udi Ndi 1 (M di vdi )
consists of feed-forward term based on Inverse Dynamics of

SLV and feedback term in which sliding mode based neural
network compensator is employed.
ydI vdI (v v v )
To design the linear controller to control the output states
(7)

, , , ,
Where, , ,
represent aerodynamic
moments, damping moment and control moment respectively.

B. Inverse Dynamic Control
The dynamic inversion methodology to formulate control
input udi requires two steps. The first step is the state and
input transformation to convert nonlinearities of the system
dynamics into its linear time invariant form:
z az bv
damp
control moment of the nominal model respectively. The

difference in actual and nominal model is represented by
term for which the moments can be written as:
aero 0aero aero ,
damp 0damp damp ,
con 0con con
The second step is the designing of the control

input v using linear control methodology. To achieve stable
tracking control, three variables of interest i.e.
ydi ( , , ) are chosen as output variables. Transform
system to get the form as shown in Equ.8, Output vector is
differentiated twice to get input vector [8]:
0damp 0con f ( damp , con )

0aero 0damp 0con f ( aero , damp , con )
In order to estimate the function
ydi ,

0
and u
di

0
f ( damp , con ) , for roll and
f ( aero , damp , con ) for pitch and yaw, adaptive Sliding mode
(14)
0aero 0damp 0con f ( aero , damp , con )
(9)
0 0 ( )
0
0

M di 0 0
0
( ) 0
0 0
0
0
0
(13)
By substituting these values, Equ. 7 can be written as:
Where,

N di 0
v d kd ( d ) k p ( d )

v d kd ( d ) k p ( d ) (12)
v d kd ( d ) k p ( d )

C. Feedback term based on Sliding mode Neural Network

In practice, it seems to be impossible to model the exact
dynamics of the system. Since the dynamic inversion control
explicitly utilizes the modeled dynamics, it is assumed
that 0aero , 0 and 0con represent aerodynamic, damping and
(8)
ydi M di Ndiudi
(11)
ydi ( , , ) , the linear controller is written as:
vDI
(10)
Where,
A. Problem formulation
In order to apply dynamic inversion to on the attitude
dynamics of launch vehicle, it is assumed that the inertial rates
p, q, r are equivalent to body rates , , neglecting the
effect of earth rotation. Therefore, the dynamic equations 4-6
can be written as:
Proceedings
Neural Network compensator (SNN) is proposed [9]. By

employing this, the control deflection computed by the
proposed controller is written as:
udi f (.)
(15)
Radial Basis Neural Network compensator is designed on

the basis of sliding surface and the surface depends on error
and its rate. RBF networks are adaptively used to approximate
the uncertain f.
f wiT hi (s)
The desired input to the system can be written as:
267
L wT (sh(s) w ) s( dt ksign(s))
Where h(s) is the Gaussian function of neural network based

on sliding surface.
s ci
hi s exp
2 i 2
Proceedings
(21)
Where is approximation error of neural network and is a

positive coefficient. The stability Criteria is defined as:
, i 1, 2,...5.
LL 0
(16)
The adaptive law is selected as
1
w sh( s)
Now we can write:
L s( dt ksign(s)) s( dt ) k s
(22)
Since approximation error is sufficiently small and
k dt
Fig. 2 SNN structure
We get
The sliding surface is defined as:[8]
L0
S C *e e
(17)
Let the nonlinear system is defined as:
x f ( x, x) gu dt
We know that
s ce e
s e ce xd x ce
(18)
D. External Guidance loop:

In order to achieve the desired orbital parameters, an
external guidance loop is incorporated in the autopilot of SLV.
The guidance scheme performs the online reshaping of the
reference attitude profiles in the exo-atmospheric phase. It
computes deviations (x, y, z) from the intended trajectory
and on the basis of these deviations, it formulates the new
reference attitude angles accordingly. The controller in turn
tracks the updated reference profiles to attain the desired orbit.
x x xd
y y yd
z z zd
gu dt ce ksign(s)
xd fx
The control input can be written as:
1
dt ce ksign(s )
u
xd fx
g
s f ( x) dt ksign(s)
(19)
Where
f f f wT h(s) w T h(s) wT h(s)

1 2 1 T
s w w
2
2
(25)
(28)
d ref
(29)
(26)
(27)
The complete block diagram of the proposed control

algorithm with external guidance loop is depicted below:
Lyapunov function is defined as:
(24)
x
z
tan 1
x
d ref
tan 1
s f ( x) f ( x) dt ksign(s)
(23)
(20)
The derivation is written as:
L ss wT w
268
Proceedings
Fig. 3 Guidance & Control architecture
IV. SIMULATION RESULTS

The performance evaluation of the proposed control
architecture was carried out on a six degree of freedom
simulation software developed in MATLAB Simulink. The
simulation result of pitch channel for the first 30 seconds of
flight including the pitch over phase under nominal flight
conditions is shown in Fig. 4.
Fig. 5 Attitude profile without Compensator with Uncertainty
Fig. 6 Attitude profile with Compensator with Uncertainty

0.6
Fig. 4 Pitch over Attitude profile with Compensator
Without compensator
With compensator
0.4
It can be seen from Fig. 4 that the proposed controller is

efficiently tracking the reference pitch angle profile. The
robustness analysis against parametric variation of the control
scheme is carried out by the incorporation of 30% random
variation in mass properties of the SLV.
Deflection [Deg]
0.2
-0.2
-0.4
-0.6
-0.8
10
15
Time [Sec]
20
25
30
Fig. 7 Comparison of Commands with Uncertainty
Fig. 5-7 show the comparison of controller performance

with and without the sliding mode neural network compensator
and their resultant outputs under parametric variations. It can
be seen that the hybrid control structure is immune to modeling
269
Proceedings
Attitude", Proceedings of International Conference on Intelligent

Autonomous Control in Aerospace, Beijing, China, (1995).
uncertainties.
The effectiveness of the proposed guidance scheme with
online trajectory reshaping is depicted in Fig. 8 that shows the
error in orbital attitude with and without guidance under
parametric variations. The guided SLV accurately attains the
desired orbit.
[4] Zhihua Qu, Robust Control of Nonlinear Uncertain Systems John Wiley
& Sons, Inc. 1998.
[5] Yuri Shtessel, James McDuffie, Mark Jackson, Charles Hall, Don Krupp,
Michael Gallaher, and N. Douglas Hendrix, "Sliding Mode Control of the
X33Vehicle in Launch and Re-entry Modes," AIAA98-4414, Proceedings
of AIAA Guidance, Navigation, and Control Conference, Boston, MA,
August 10-12, 1998
[6] Ito D, Ward DT and Valasek J. Robust dynamic inversion controller
design and analysis for the X-38. In: AIAA conference on guidance,
navigation and control, Canada, 69 August 2001, paper no. AIAA-20014380.
[7] David B. Doman and Anhtuan D. Ngo Dynamic Inversion Based
Adaptive/Reconfigurable Control of the X-33 on Ascent Journal of
Guidance Control and Dynamics, 2002.
Fig. 8 Injection Altitude with uncertainties
Fig. 9 shows the pitch profile for the whole mission from
lift-off to orbit injection including online attitude reshaping
from guidance scheme to attain the desired orbital parameters.
[8] Abhijit Das, Frank L. Lewis and Kamesh Subbarao (2011). Sliding Mode
Approach to Control Quadrotor Using Dynamic Inversion, Challenges
and Paradigms in Applied Robust Control, ISBN: 978-953-307-338-5.
[9] Feng G (1995), A compensating scheme for robot tracking based on
neural networks. Robot Auton Syst 15(3):199206.
Pitch Angle [deg/sec]
100
Actual
Reshaping
Reference
50
-50
100
200
300
400
500
600
Time [sec]
Fig. 9 Pitch profile with uncertainties
V. CONCLUSION
An adaptive and robust control strategy for SLV autopilot
is presented in this paper that is based on inverse dynamic
control aided with a sliding mode neural network compensator.
Lyapunov stability criterion is utilized to ensure the stability of
the control algorithm. The simulation results prove that the
proposed controller not only guarantees stability but also
efficiently tacks the reference attitude profile and the steering
commands from the guidance loop even under parametric
variations. The guidance scheme performs the online reshaping
in the exo-atmospheric phase of flight in case of deviations
from the intended trajectory to meet the desired orbital
parameters accurately.
REFERENCES
[1] Uzair Ansari. "Hybrid Genetic Algorithm fuzzy rule based guidance and
control for launch vehicle", 2011 11th International Conference on
Intelligent Systems Design and Applications, 11/2011
[2] Ansari, Uzair, Saqib Alam, and Syed Minhaj un Nabi Jafri. "Trajectory
optimization and adaptive fuzzy based Launch Vehicle attitude control",
2012 20th Mediterranean Conference on Control & Automation (MED),
2012.
[3] Filho, W. C. L., "Application of Adaptive Control Sounding Rocket
270
Proceedings
Design and Development of Low cost Motor Drive

for Hub Wheel based Electric Vehicles
Syed Wasif Ali Shah
Muhammad Asim
SUPARCO-Karachi-Pakistan
wasif_iiee@yahoo.com
SUPARCO-Karachi-Pakistan
m_asim2000@hotmail.com
AbstractRobotic mobility platforms are serving a wide

range of civil, commercial, domestic and military applications.
This paper describes the design and development of low cost
motor drive, applicable for wide range of permanent magnet DC
(PMDC) motors for electric vehicles and especially for the hub
wheel motors as they exhibit a distinctive frequency behavior.
This controller comprises of 3 units: First, a logic unit providing
control signals; Second, a sensing unit for over current and
batteries protection: third, a high power switching unit for
driving motors. Though applicable to wide range of unmanned
ground vehicles (UGVs) but here it is implemented in a
differential drive autonomous UGV built on hub motor wheels
with a speed of 1.88 m/sec, endurance of 2 hours and payload
capacity of more than 100 kg. Features, performance and
development cost of the drive in comparison with the similar
commercially available drives is also discussed.
Keywords Mobility platforms; UGV; Hub Wheel; Motor
Drive;
Particularly we have developed a UGV shown in fig: 1 with

the dimensions of 72 cm in length, 91 cm in width and 20 cm
in height, having weight of 90 kg and payload capacity of
more than 100 kg build on the Hub motorized wheels. This
UGV is primarily designed for autonomous transportation
with GPS navigation with maximum speed of 1.88 m/sec. The
UGV has the differential drive mechanism with zero turning
radius. The motor drive has shown the ability to smoothly
control the hub motors torque, rpm and direction. The
advantage of hub wheel motor in building the UGV is its
mechanical simplicity and installment, compactness giving
more room to other components, less complex structure of
motion transmission eliminating axels and drivelines, reduced
weight giving more efficiency and endurance and ability of
high torque at low rpm [10]. The implemented hub motor has
the following major parameters given in tab-1.
Table 1: Hub Motor Parameters
I.
INTRODUCTION
Unmanned ground vehicles (UGVs) are the mobility platforms

that have enormous civil, commercial, military, domestic and
urban applications [1][2] ranging from rescue, surveillance,
reconnaissance, mine detection, fire fighting, material
handling, agriculture, handicapped assistance, border security,
special weapons and tactics (SWAT) robot, manipulation,
exploration, automation, transportation, handling in dangerous
environment and many others. UGVs may be based on
different propulsion systems like engine based, electric or
hybrid drive mechanism [3]. Similarly different applications
and features require different traction and locomotion systems
[4] depending upon the required speed, tracks, obstacle
traversing, energy efficiency, load bearing capacity,
mechanical and control complexity and many others. To meet
different applications and fulfill different tasks through UGVs,
varieties of motorized actuation systems are available and
applied; also different mechanical structures and their level of
complexity have different actuation systems [5].DC motors
amongst them are most widely used actuator in electric
vehicles due to their low cost, simple and economical control
and weight to performance ratio [6].Varieties of DC motor
drives are available having different options, ratings and
control interfaces [7][8][9]. The motor drive described here
has the ability to drive wide range of permanent magnet
brushed DC motors with power rating of up to 1.8 KW.
271
Parameter
Value
Power
Voltage
RPM
Radius
250W
24 Vdc
120
6 in
Figure 1: The differential drive UGV build with Hub motorized wheels
Proceedings
The objective of the design is to develop a high current low

cost PMDC motor driver primarily for heavy robot building
and applications since available commercial drives are either
very expensive or inadequate for large applications. The
presented motor drive can be used for a variety of PMDC
motors. Though primarily developed for UGVs but suitable
for wide range of industrial and commercial applications like
blowers, pumps, valves etc and other actuators.
II.
DRIVE DESIGN
Figure 3: The logic unit accepting virtually all compatible signal sources.
The design of the motor drive is consisted of three main units

as shown in block diagram fig-2. Despite these units a voltage
regulation circuitry is added to provide stable power to logic
unit and optional 12V fan. Its a high efficient switching
regulator which also supplies off the board power through
connector to external logic controller eliminating the need of
another power source required for microcontrollers, Radio
Control (RC) receivers and other signal sources.
A. Logic Unit
The logic layer bridges the input signal to the power unit to
drive motor and simultaneously gets input from sensing unit to
monitor current flow through the drive and battery voltage
level. This unit accepts the TTL level pulse width modulated
(PWM) and logic signals and translates to the control signals
for power unit for forward, reverse direction, speed, brake and
enable/disable commands. The PWM and intelligence signals
are provided externally. The drive can be interfaced with any
microcontroller, timing circuitry or simply with a computer
easily. A variety of other signal sources can also be interfaced
by adding the converter circuitry to PWM like it can be
interfaced with RC receivers to operate wirelessly with RC
radio, serial, I2C and other interfaces as shown in fig-3. The
logic unit is electrically isolated from the power unit to avoid
any damage to the control circuitry.
B. Sensing Unit
As shown in the block diagram fig: 2, sensing unit monitors
the current being drawn from the power source to the motor
through power unit as well the voltage level of the attached
batteries. Whenever the vehicle stucks in the rough terrain or
passing through the uneven surfaces, it leads to draw the extra
amount of current or even stall conditions which may damage
the motor and the power unit of the drive. So it ensures the
safety by limiting the current to the motor as adjusted by the
user.
The hall sensor is employed to sense the current. Upper
maximum limit can be set in this unit above which it gives the
signal to the logic layer which forces all the output levels to
zero regardless of the input signals and turns off the operation.
It also checks the batteries voltage level and perform the same
action to turn off the drive if it falls below the certain level to
avoid damage to the system and batteries life. A graph of trip
voltages against the adjusted peak current is shown in fig: 4.
C. Power Unit
High power fast switching unit is the last stage of the drive. It
receives direction, speed and brake commands and 13-50 volt
power and can pulsate it at max up to 16 KHz rate. In result it
drives the motor with max of 160 A of continuous current.
The block diagram of the power unit is shown if fig-5.
Figure 2: The block diagram of the motor drive consisting of three main units.
Figure 4: Trip voltage against the adjusted peak current.
272
The FET driver accepts signals from the logic unit and drives
the FET bridge circuitry. It can be seen that block diagram is
in full bridge configuration. The bridge is made of the NChannel power MOSFETs as shown in one of the four legs in
fig: 6. FET driver has built in voltage booster. It outs four
signals to drive the full bridge circuitry and accepts the
modified TTL signals allowing a variety of signal sources to
drive. The FET bridge driver provides voltage above the
positive voltage source in to the gate of high side FETs
through H1 and H2 to turn on the MOSFETs. FETs usually
are sensitive to voltage levels. A higher or lower level of
voltage may damage it. Therefore Zener diodes are added in
the circuit as shown in fig: 6 to protect the gates of MOSFETs.
Two Zener diodes are used to clip the upper and lower level of
voltages. Also these diodes clip the fast charging and
discharging gate capacitance due to transistors switching.
Proceedings
Figure 6: Power unit bridge circuitry showing of the leg.
with the electrolytic capacitors and snubber circuits. Optional

fan output of 12 V is provided to connect the cooling fan to
manage the dissipated heat through the MOSFETs due to high
current.
III.
HUB MOTOR FREQUENCY RESPONSE
The most commercially available motor drives dont

provide the option to change the frequency of the switching
power to the motor. They change polarity to alter the direction
and pulse width to vary the speed at fixed frequency as it
usually not required, whereas the hub motor has a distinct
frequency behavior as compared to other PMDC motors. It
consumes different power at different switching frequencies.
Under no load conditions, the motor draws higher current at
low switching frequencies which even increases till 1.5 kHz
but drastically decreases up to 4 kHz and then gradually
further decreases and almost becomes steady at higher
frequencies as shown in fig: 7.
The drive features the variable frequency (to be provided
by the external controller) up to 16 kHz depending upon the
number of MOSFETs used in the power unit. Higher the
numbers of MOSFETs limit the maximum switching
frequency.
Figure 5: The high switching MOSFET power unit
Single to several MOSFETs can be used in one leg of the

bridge depending upon the motor current requirement. A
single MOSFET used here can bear up to 40 A of continous
current hence four can stand the load of 160 A continous. This
drive can be used in parallel configuration to drive the even
more higher power and current motors and other loads upto
300 A without much worrying about dissipated heat
management. A small resistor is added between the FET and
driver to limit and control the possible large gate current flow
due to multiple MOSFETs used in single leg as shown in fig:
6. Schottkey diodes are placed in parallel with the gate
resistor, starts to conduct when the FETs are turning off by
enhancing its time and eliminates the possiblilty of shootthrough situation. Motors and other inductive loads creates
problem for the drives due to voltage spikes and noise of
brushes and commutators. Transient voltage suppressores
(TVS) are sued as the zener diodes to safely handle the high
current voltage spikes coming from motors. TVS are
connected parallel to battery voltages to clip the spikes across
the board. Additional filteration of the power source is done
273
IV.
DEVELOPMENT COST
Since the drive is designed to achieve a low cost dc motor

controller while differential drive UGVs require two parallel
motors to be excited. Hence two drives are employeed in our
plateform. The drive is build from the commercially and easily
available components in the local market hence making it
serviceable and replaceable. The total development cost of the
motor drive is listed in the tab: 2.
Figure 7: Frequency response of the hub motor showing current variation at

different switching frequencies
LinoGuzzella, Antonio Sciarretta Vehicle propulsion systems

Introduction to Modeling and Optimization, 2nded.
[4] L Bruzzone and G. Quaglia Review Article: Locomotion systems for
ground mobile robots in unstructured environment Mech:sci:, 3,49-62,
2012, www.mech-sci.net/3/49/2012/
[5] William Brown Improving the electromechanical performance of Robot
Arms on Unmanned Ground Vehicles GRRC Technical report 2010-01,
12/20/2009
[6] Nasser Hashemnia and Behzad AsaeiComparative study of using
different electric motors in the electric vehicles proceedings of the 2008
international conference on electrical machines, paper ID 1257.
[7] Roboteq, Inc., Brushed DC motor controller, Online:
http://roboteq.com, 2015.
[8] Robot Power. Robot power products, Online: http://robotpower.com,
2015.
[9] VEX
Robotics.
Victor
Motor
Controller.
Online:
http://vexrobotics.com, 2015.
[10] Machine Design. A look at the advantages of hub motors as well as
disadvantages, Online: http://machinedesgin.com, 2015.
[3]
Table 2: Costing of the motor drive.

Component
Drive PCB
Drive Components
Development
Total
Appromimate
Cost
$10
$48
$7
$65
Table 3: Specifications of the motor drive

Specification
value
Supply voltage
Output Current (continuous)
Output Current (surge)
Switching Frequency
Weight
MOSFETs
On Resistance
Cooling
Dimension (mm)
13V to 50V
160A
>400
Upto 16 kHz (with 4 MOSFESTs)
100 g (without fan)
16 (4 per leg)
.0026 ohm max at 25C
Through optional 40 CFM 12 volt
fan
115(L)X80(W)X35(H)
Parameters and features of the drive are given in tab: 3. The

drive has competitive features and performance if compared
with similar commercially available PMDC motor drives
whose prices range from $200-$400 [7][8][9]. The drive is
thoroughly tested with different motors and loading conditions
for different applications. It conforms to the specifications
especially under harsh and open environment of robotic
mobility platforms for long durations. Different features can
also be added in this design with minimal marginal cost which
otherwise incurs heavy cost if procured.
V.
CONCLUSION
The paper discussed the design and development of low cost

PMDC brushed motor drive. Particularly implemented in
differential drive UGV with heavy payload capacity, build
with hub motorized wheel showing distinctive frequency
response. The developed controller consists of three units:
logic unit, sensing unit and high switching power unit. The
drive can be driven virtually with any microcontroller and
other signal sources like PC and wirelessly with RC signals. It
senses the current being drawn and the battery voltage levels
and further can be enhanced to monitor other motor
parameters like temperature. Features and development cost of
the drive are also discussed and compared with similar drives
available commercially. The drive is very economical with
easy service and replacement.
VI.
[1]
[2]
Proceedings
REFERENCES
V. Sezer, C. Dikilita, Z. Ercan, H. Heceoglu, A.Buner, A.Apak,

M.Gokasan and A. Mugan Conversion of a conventional electric
automobile into an unmanned ground vehicle (UGV) IEEE
international conference on Mechatronics, pp. 564-569, (2012)
A. Bouhraoua, N.Merah, M. Al-Dajani and M. Elshafi Design and
development of an unmanned ground vehicle for security applications
7th International symposium on mechatronics and its applications, pp.
1-6, (2010)
274
Proceedings
Preparation, Structure and Dielectric/Piezoelectric

Properties of BiScO3-PbTiO3-Pb(Mn1/3Nb2/3)O3 High
Temperature Piezoelectric Ceramics
Shahid Karim1, Jinting Tan1, Adil Murtaza2, Shahid Sultan2, Khalida Kareem2, Muhammad Nouman Shaikh3
1
Electronic Material research laboratory & International Center of dielectric research, School of Electronics and Information
Engineering, Xian Jiaotong University, Xian 710049, China
2
School of Sciences, Frontier Institute of Science and Technology, MOE Key Laboratory for Nonequilibrium Synthesis and
Modulation of Condensed Matter, State Key Laboratory for Mechanical Behaviour of Materials, Xian Jiaotong University, Xian
710049, China
3
Xian Jiaotong University, Xian 710049, China
AbstractWith the rapid growth of science and technology,
more piezoelectric devices should be used in high-temperature
environment. Commercially using PZT piezoelectric ceramics
have a low curie temperature (<150C). So the present study is
focused on synthesis and characterization of high Curie
temperature piezoelectric ceramics. As a kind of excellent high
Curie temperature piezoelectric ceramics BiScO3-PbTiO3 (BSPT) system at MPB reveals high Curie temperature piezoelectric
properties. Pb(Mn1/3Nb2/3)O3-PbTiO3 (PMnN-PT) system
ceramics have excellent piezoelectric properties at MPB with low
Curie temperature (150C). 0.1PMnN-BSPT ceramics show the
optimal piezoelectric properties with d33=200pC/N, kp=0.40,
Tc=338C, Ec=22.75kV/cm and Pr=21.85C/cm2. The 0.05PMnNBSPT ceramics indicate the optimal piezoelectric properties with
d33=250pC/N,
kp=0.54,
Tc=373C,
Ec=14.63kV/cm and
Pr=11.51C/cm2. Also 0.06PMnN-BSPT system with 64mol.% PT
indicates the optimal piezoelectric properties with d33=210pC/N,
kp=0.49, Tc=328C, Ec=20.21kV/cm and Pr=8.35C/cm2. So the
introduction of PMnN reduces the curie temperature; 10mol.%
of PMnN indicates the hard behavior as a function of reducing
the piezoelectric properties; 6mol.% of PMnN indicates the
soft behavior with increasing the piezoelectric properties.
Index TermsMorphotropic phase boundary (MPB), Curie
temperature,
Piezoelectric
Properties,
High-temperature
piezoelectric ceramics, Binary system ceramics
I. INTRODUCTION
Piezoelectric ceramic is a kind of important functional
ceramic material, which can realize the conversion between
mechanical energy and electric energy which are widely used
in manufacturing, micro displacement transducer and
ultrasonic generator etc. With the development of science and
technology, a lot of electrical and electronic equipment which
are used in high temperature piezoelectric devices increased
the demand. Study on the high temperature, with excellent
piezoelectric properties of piezoelectric ceramic materials
have become a hot research topic of functional ceramics
industry. The commercial applications of lead zirconate
titanate (PZT) in piezoelectric ceramics range of Curie
temperature system is 250-380C, and the temperature is
limited to 150C or less, which is far less than high
temperature piezoelectric devices. For a long time, a lot of
piezoelectric crystal materials due to the use of high

temperature and excellent piezoelectric properties have
become an important part of high temperature piezoelectric
applications [4]. Therefore high electromechanical power
applications requires such ceramics with better electrical
properties such as the piezoelectric strain constant (d33>200
pC/N), low dielectric loss (tan) mechanical quality factor
(Qm>1000), planar electromechanical coupling factor (kp> 0.5)
[1-3]. Lead niobate (PbNb2O6) tungsten bronze structure was
first discovered, it has tetragonal tungsten bronze structure [5].
This system has large anisotropy piezoelectric coefficient,
quality factor, low and high Curie temperature (570C) is not
easy to depolarization, therefore particularly suitable for the
manufacture of high temperature transducer. But pure lead
niobate piezoelectric activity difference exists, the
electromechanical coupling coefficient is low, and is not easy
to sinter, and its high temperature ferroelectric phase at room
temperature is not stable. There is irregular grain growth and
phase transformation during sintering and cooling process
caused by the large volume change, so it is difficult to prepare
the excellent performance of the pure lead niobate
piezoelectric ceramic materials [6]. Through the substitution
of bismuth layer structure piezoelectric ceramic modification
can improve its piezoelectric activity. Korzunova [5] in the
Bi3TiNbO9 system, using Ti4+, Ta6+ or W6+ instead of B in the
Nb5+, using K+ to replace Bi3+, the piezoelectric constant 5
pC/N is increased to 24 pC/N.
II. EXPERIMENTAL PROCEDURE
Mixture grade (>99.9% purity) powders of Bi2O3, PbO,
Sc2O3, TiO2, MnCO3, and Nb2O5 were used to make xPbTiO3(0.9-x)BiScO3-(0.05 to 0.15)Pb(Mn1/3Nb2/3)O3, (x=0.64). The
mixture of MnCO3 and Nb2O5 was ball-milled in a
polyethylene jar for 4h by high purity tetragonal zirconia
polycrystalline (3Y-TZP) ball media (Tosoh Co., Tokyo,
Japan). The mixed slurry was dehydrated and condensed for
calcination and phase formation at 800C, 850C and 900C
for 5h. The Mn1/3Nb2/3O2 phase was confirmed by X-ray
diffractometer (XRD) after calcination, to obtain fine particles
the calcined piece was ball milled again for 4 h. After
preparation of Mn1/3Nb2/3O2 powder, PbO, Bi2O3, Sc2O3, and
275
TiO2 were mixed with it and ball milled to make final

compositions. Pellets (12.7x3mm) were compressed under a
pressure of 100 MPa by using a hydraulic press. Pressureless
sintering was performed in a high purity alumina crucible at
1100C, 1150C and 1200C for 2h. To prevent from reaction
with alumina crucible, the specimens were put on a Pt-foil
inside crucible. After sintering, apparent densities were
measured by the Archimedes method.
Electrode formation was done by sputtering silver on the
lapped surfaces of pallets. By applying dc field of 3.5 kV/mm
for 10 min, electroded specimens were poled at 120C in
silicone oil. Dielectric properties were acquired by measuring
the capacitance and loss using LCR meter (HP model 4284).
To determine Curie temperature, capacitance measurements
were completed as a function of temperature in an automated
temperature controlled furnace linked with a computer for data
acquirement. The piezoelectric coefficient (d33) was noted from
one-day aged samples by using Berlincourt-d33 meter.
Electromechanical coupling factor (kp) and mechanical quality
factor (Qm) were obtained by using an impedance analyzer
(Model HP4194) built on the IEEE standards [12]. Each
experimental step is given in Fig. 1.
Make the composition of
Mn1/3Nb2/3O2 by the
molecular weight of
MnCO3, and Nb2O5
Proceedings
seen from the angle 44 to 46 at 200 peak, all samples show

tetragonal and rhombohedral phases.
Fig. 2. Powder X-ray Diffraction Analysis of Samples
piezoelectric coefficient d33

was recorded using a
Berlincourt d33 meter,
dielectric properties were
obtained using an LCR meter
Ball milled,4h
The electroded
specimens were poled in
o
silicone oil at 120 C by
applying a d.c. field of
3.5 kV/mm for 10 min
Calcinate it at
1000'C for 5 hours
Mix the powders of Mn1/3Nb2/

3O2, PbO, Bi2O3, Sc2O3 and
TiO2 and take six compositions of
5%, 6%, 8%, 10%, 12% and 15%
Ball milled,4h
Calcinate each sample
at different temperature
800'C, 850'C and 900'C
respectively
XRD
Blend the powders by
mixing with PVA+
Water and take the
specific grain size of
powders to make pallets
Fig. 3. XRD of the samples of x=5, 6, 8, 10, 12, 15% respectively

sintered at 1150C
silver was pasted for

the electrode
formation
XRD
Make the pallets
shape smooth and
then measure the
apparent densities of
each sample
Sintered each
sample at different
temperatures
1150'C, and 1200'C
Fig. 1. Flow chart of experiment

A. Sintering behavior and microstructures
Fig. 2, shows Powder XRD for the samples 5, 6, 8, 10, 12
and 15% of PMnN calcined at 800C, 850C and 900C. It is
clearly shown at 8% of PMnN the XRD pattern is smooth.
Also it is studied that all the samples show tetragonal and
rhombohedral phases. Fig. 3, shows sintered XRD for the
samples 5, 6, 8, 10, 12 and 15% PMnN which are sintered at
1150C. There is clearly shown at all doping percentage of
PMnN the XRD pattern is smooth. Also it is seen from the
angle 44 to 46 at the peak of 200 of all the samples show
tetragonal and rhombohedral phases. Fig. 4, shows sintered
XRD for the samples 5, 6, 8, 10, 12 and 15% PMnN which are
sintered at 1200C. There is clearly shown at all doping
percentage of PMnN the XRD pattern is smooth. Also it can be
Fig. 4. XRD of the samples of x=5, 6, 8, 10, 12, 15% respectively

sintered at 1200C
The microstructures of the composites were studied by

Scanning electron microscope (SEM). The SEM micrographs
of the samples (a) 5%; (b) 6%; (c) 8%; (d) 10%; (e) 12% and
(f) 15% of PMnN sintered at 1150C for 2h are shown in Fig.
5. They showed different microstructures distinctly; grain size
of each sample is (a)3.6; (b)2.8; (c)3; (d)3.2; (e)3.7; and (f)2; in
micrometers respectively and grain boundaries of sample
(c)8% are more clear and significant. The SEM micrographs of
the samples 6% of PMnN sintered at (a) 1100C; (b) 1150C
and (c) 1200C respectively for 2h are shown in Fig. 6. They
showed different microstructures distinctly; grain size of each
276
sample is (a)1.6; (b)2.8; (c)4; in micrometers respectively. The

SEM micrographs of the samples 8% of PMnN sintered at (a)
1100C; (b) 1150C and (c) 1200C respectively for 2h are
shown in Fig. 7. They showed different microstructures
distinctly; grain size of each sample is (a)1.6; (b)2.6; (c)4.2; in
micrometers respectively. The SEM micrographs of the
samples 12% of PMnN sintered at (a) 1100C; (b) 1150C and
(c) 1200C respectively for 2h are shown in Fig. 8. They
showed different microstructures distinctly; grain size of each
sample is (a)1.6; (b)3.8; (c)5; in micrometers respectively. It is
clearly seen that in all samples of Fig. 6, 7 and 8 there is no
significant change in the grain size of samples sintered at
1100C but the grain boundary is clear by increasing the
content of PMnN. Each SEM microstructure of the sintered
body did not show any second phase and with increasing
temperature, grain size was increased and grain boundaries are
more clear and significant. Nevertheless the microstructures of
the every composition are uniform.
Proceedings
Fig. 7. SEM microstructures of Samples 8% of PMnN sintered at (a)

1100C, (b) 1150C and (c) 1200C respectively

B. Bulk Density
The density of undoped BSPT and from 5 to 15 mol.% of
PMnN of all the samples are given as shown in Fig. 9. It is
clearly seen that the density of sample at 5% is more than
others.
C. Dielectric Properties
Fig. 10, illustrate dielectric constant as a function of
temperature for BSPT-(5-15%)PMnN ceramics at different
frequencies (1kHz~100kHz). It is difficult to see that the
0.64PT content of ceramic samples with diffuse phase
transition and frequency dispersion phenomenon, the width of
dielectric peaks in the vicinity of the characteristic temperature
change with the increase of frequency, the dielectric constant,
dielectric peaks shifted to higher temperature. With the
increase of PMnN content, the dielectric peaks become sharp
and frequency dispersion phenomenon gradually disappeared.
Fig. 11, is the medium temperature of BSPT-PMnN ceramics
samples in the 1 kHz spectrum. It can be seen that at different
content of PMnN, sometimes the Curie peak becomes sharp,
dielectric constant, characteristic temperature are sometime
increased and sometime decreased. At 5% of PMnN there is
highest temperature and highest dielectric constant in all
compositions as shown in Fig. 10.
Fig. 5. Sample (a) 5%; (b) 6%; (c) 8%; (d) 10%; (e) 12% and (f) 15%
of PMnN sintered at 1150C respectively

D. Piezoelectric Properties
The electromechanical coupling factor kp was obtained by
using an impedance analyzer. Electromechanical properties
are measured from the admittance spectra at low field ac drive
are shown in Fig. 12, as a function of PMnN compositions.
Near the MPB boundary, 64% PT the sample of 5 and 15% of
PMnN there is a significant increase in kp factor, at 12% there
is decrease in kp factor as clearly shown in Fig. 12. However
the consistency of electromechanical properties lowers
considerably while driving at high-temperature.
According to high Curie temperature and the coercive field
of BSPT, this composition is predicted as high power
piezoelectric materials, if we can increase the mechanical
quality factor Qm. Near the MPB boundary, 64% PT the sample
of 5 and 10% of PMnN there is clearly increment in quality
277
factor (Qm) as shown in Fig. 13. So, it has been seen that
replacement of Pb(Mn1/3Nb2/3)O3 is meaningful to improve
mechanical quality factor Qm of BS-PT.
All the properties show highest magnitude at compositions
near the MPB boundary at 64% PT. There is significant
increase in the magnitude of d33 for the samples of BSPT and
other compositions of 5, 6, 8, 10, 12 and 15% of PMnN
respectively. At these compositions the magnitude of d33 are
320, 250, 210, 220, 200, 74 and 160pC/N respectively as
shown in Fig. 14. Dielectric and piezoelectric properties of all
samples are given in table I.
Proceedings
Fig. 11. Dielectric constant as a function of temperature for BSPT-PMnN

ceramics at frequency of 1 kHz
Fig. 9. Apparent densities at different temperatures 1150C and 1200C

Fig. 12. Electromechanical Coupling Factor at different doping percentages 5,
6, 8, 10, 12 and 15% of PMnN
Fig. 13. Mechanical Quality Factor at different doping percentages 5, 6, 8, 10,

12 and 15% of PMnN
Fig. 10. BSPT-PMnN system ceramics dielectric temperature under different

frequency spectrum
Fig. 14. Piezoelectric constant for all samples at doping %age of 5, 6, 8, 10, 12
and 15 of PMnN
278
TABLE I. Dielectric and Piezoelectric Properties with Curie temperature

Systems
d33 (pC/N)
Tc (C)
kp
Qm
max
5%
250
373
0.54
30
1018
12820
6%
210
328
0.49
15.6
761
4874
8%
220
302
0.42
27.06
1149
5104
10%
200
338
0.40
59.60
931
9809
12%
74
286
0.35
14.5
764
3169
15%
160
326
0.50
13.38
927
10248
IV. EFFECTS OF ANNEALING ON DIELECTRIC AND

PIEZOELECTRIC PROPERTIES OF BSPT-PMNN CERAMICS
Microstructure of ceramics sturdily depends on annealing
process and it defines the ceramics physical performance.
Density, homogeneity and grains size in the annealing steps are
significant to attain the preferred electrical properties [7]. It
will influence the crystallographic phase, microstructure and
electrical properties of the resultant thin films [8-10]. The
temperatures of pre-annealing and annealing were determined
by the results of the thermal analysis and XRD, respectively.
E. Ferroelectric Properties
Polarization is saturated PS at high electric fields and there
is no more motion of the domain walls. Area of the hysteresis
curve shows heat density generated in the material at one
cycle. It is clearly seen that the 10% of PMnN has high
polarization value as well as electric field as shown in Fig. 15.
If the residual magnetization is high when magnetic
intensity will zero, it means that material have strong value of
magnetization. If the value of residual magnetization is less, it
means it will not retain the magnetization for long time after
removing the magnetic intensity. The results of remanent
polarization and coercive field were given in tab. II.
When the electric field began to decrease because of the
crystal stress domain deviate from the direction of
polarization, the electric field is reduced to zero, still remain in
the polarization direction of ferroelectric domains, thus the
macro reflects the remnant polarization, so that the material is
polarized. When the electric field is reversed, the remnant
polarization disappears, so the coercive field can be regarded
as the minimum field strength of the domain.
Fig. 15. Polarization of all compositions with 5, 6, 8, 10, 12 and 15% of PMnN
TABLE II.
System
5%
6%
8%
10%
12%
15%
Remanent Polarization and Coercive Field

Remanent Polarization
(Pr/C/cm2)
11.50
8.36
10.99
21.85
9.42
9.82
Coercive Field
(kV/cm)
14.63
20.21
11.03
22.75
13.81
19.78
Proceedings
A. SEM Analysis After Annealing

The microstructure of the compositions after annealed was
investigated by SEM. The SEM graphs of the samples (a) 5%;
(b) 6%; (c) 8%; (d) 10%; (e) 12% and (f) 15% of PMnN
respectively annealed at 800C for 2h are shown in Fig. 16.
SEM microstructures of the annealed body are not showing any
second phase. They have clearly displayed different
microstructures; grain size of each sample is (a)3.2; (b)2.8;
(c)7.2; (d)4; (e)2.2; and (f)2; in micrometers respectively and
grain boundaries of sample (c)8% are more clear and
significant. They showed various microstructures clearly with
annealed temperature of sample 8% of PMnN, grain size was
increased and grain boundaries of sample (a) and (d) are also
clear and significant. Nevertheless the microstructures of the
every composition are uniform.
B. Piezoelectric Properties After Annealing
The piezoelectric coefficient d33 was noted from one-day
aged samples by a Berlincourt d33 meter (IAAS ZJ-2, Beijing,
China). All the properties displayed highest magnitude of
compositions near MPB boundary that is 0.64PT. As compared
to BS-PT ceramic, there is noticeable decrease in the
magnitude of d33 [11]. Also there is noticeable increase in the
magnitude of d33 after annealing for the samples of BSPT and
other compositions of 5, 6 and 12% of PMnN respectively. At
these compositions the magnitude of d33 are 490, 320, 270,
230, 210, 120 and 165pC/N respectively as shown in Fig. 17.
The electromechanical coupling factor kp was measured
using an impedance analyzer (Model HP 4194) based on the
IEEE standards [12]. Electromechanical resonance properties
determined after annealed at 800C for 2h from the admittance
spectra at low field ac drive are shown in Fig. 18, as a function
of PMnN composition. Near the MPB boundary, 64% PT the
sample of 10 and 12% of PMnN there is a bit increase in kp
factor, at 5 and 15% there is decrease in kp factor as clearly
shown in Fig. 18.
The factor Qm was calculated by using an impedance
analyzer. As a consequence of the higher amount of coercive
field and Curie temperature of BS-PT, this composition is
predictable as high power piezoelectric material, if we can
increase the value of quality factor Qm. Near the MPB
boundary, 64% PT the sample of 5, 10, 12 and 15% of PMnN
there is clearly increment in quality factor (Qm) as shown in
Fig. 19. So, the replacement of Pb(Mn1/3Nb2/3)O3 is effective to
improve the mechanical quality factor Qm of BS-PT system.
279
C. Ferroelectric Properties After Annealing

There is a dependency of the piezoceramic properties on
the achieved polarization level that affects the uniformity of the
results due to edge effects, achieved from length poled
resonators, long bar and the standard in-plane poled shear plate
and the thickness poled resonators; like the thin disk or plates.
It is studied that all the results for polarization have increased
significantly after annealing temperature. There clearly shown
in Fig. 20, that every sample has a large difference in electric
field and polarization. And the saturation point of sample at 8%
of PMnN is very small and the curve is steep. The saturation of
6% of PMnN has larger value than others.
It has been studied that the value of remnant polarization
and coercive field were increased conspicuously after
annealing temperature. And for the sample 5% of PMnN there
is no difference in the value of coercive field before and after
annealing. All the extracted data for polarization is mentioned
in tab. III.
Proceedings
Fig. 17. Piezoelectric Constant after annealing temperature of 800C
Fig. 18. Electromechanical coupling factor of composition of 0, 5, 10, 12 and

15% of PMnN
Fig. 19. Mechanical Quality Factor of compositions of 0, 5, 10, 12 and 15% of

PMnN
Fig. 16. SEM micro graphs are shown of Samples (a) 5%; (b) 6%; (c) 8%; (d)
10%; (e) 12% and (f) 15% of PMnN respectively after annealing
Fig. 20. Polarization Characteristics of 5, 6, 10, 12 and 15% of PMnN
280
TABLE III.
System
5%
6%
10%
12%
15%
BSPT
Remnant Polarization and Coercive Field after annealing

Remnant Polarization
(Pr/C/cm2)
23.30
24.93
25.63
17.35
21.29
33.98
Proceedings
writing. I would not have accomplished this without their

constant support and feedback; they made it a thorough
learning and pleasant experience for me. I admire the guidance
and assistance provided by research associates and fellows.
Coercive Field
(kV/cm)
-14.66
-17.02
-20.55
-17.01
-21.17
-17.33
REFERENCES
V. CONCLUSION
BSPT-PMnN dense ceramic samples were prepared by the
columbite pre-production process. XRD results show that, at
the appropriate sintering temperatures, two systems were
obtained with pure perovskite structure. In systems of
0.05PMnN-BSPT,
0.06PMnN-BSPT,
0.08PMnN-BSPT,
0.1PMnN-BSPT, 0.12PMnN-BSPT and 0.15PMnN-BSPT
when the content of PT is 64% is the tetragonal-phase
existence of morphotropic phase boundary region. SEM
images show that the ceramic samples grain sizes are in
uniform distribution at the appropriate sintering temperatures,
the grain size is about 1m. When the content of PMnN is
low, the PMnN-BSPT system shows the characteristics of
typical relaxor ferroelectrics, with the increase of PMnN
content, gradually transition from ferroelectric relaxor to
normal ferroelectrics. Due to low content of PMnN, PMnNBSPT system does not show obvious relaxation
characteristics. With the increase of PMnN content, the
remnant polarization Pr, first increased and then decreased.
The coercive field system gradually increases with the
increase of PMnN content. The piezoelectric properties of the
optimal compositions of 0.1PMnN-BS-PT and 0.05PMnNBSPT system in the morphotropic phase boundary region are:
d33=250pC/N, kp =0.54 and the Curie temperature of
Tc=373C. The piezoelectric properties of the optimal
composition of 0.05PMnN-0.95BSPT system in the
morphotropic phase boundary region are better than other
optimal compositions.
After annealing at 800C for 2h, the resident strain can be
quited and the order degree of lattice may be higher than that
without annealing process. Annealing process may affect the
domain size and domain switch, domain size will be larger
after annealing process. The value of Qm increase evidently
after annealing process.
ACKNOWLEDGMENT
[1] Ueha, S., Ultrasonic motors: theory and applications, Vol. 29.
1993: Oxford University Press, USA.
[2] Chen, H., The influence of Pb(Mg1/3Nb2/3)O3 content on the
piezoelectric and dielectric properties of PMNN-PZT system
near the MPB. in Applications of Ferroelectrics, 2002. ISAF
2002. Proceedings of the 13th IEEE International Symposium
on. 2002. IEEE.
[3] DU, H. l., Effect of sintering temperature and composition on
microstructure and properties of PMS-PZT ceramics,
Transactions of Nonferrous Metals Society of China, Vol. 16:
pp.165-169, 2006.
[4] Weis, R. and T. Gaylord, Lithium niobate: summary of
physical properties and crystal structure, Applied Physics A,
Vol. 37(4), pp. 191-203, 1985.
[5] Korzunova, L. V., and L. A. Shebanov, New perovskite-like
high temperature ferroelectrics, Ferroelectrics, vol. 93.1 pp.
111-115, 1989.
[6] B. Frit and J. P. Mercurio, The crystal chemistry and dielectric
properties of the Aurivillius family of complex bismuth oxides
with perovskite-like layered structures, Journal of Alloys and
Compounds, vol. 188, pp. 27-35, 1992.
[7] M. Ghasemifard, Dielectric, piezoelectric and electrical study
of 0.65PMN-0.20PZT-0.15PT relaxor ceramic, The European
Physical Journal Applied Physics, Vol. 54, No. 2 , pp. 2070120707, 2011.
[8] Laughlin, Brian, Jon Ihlefeld, and JonPaul Maria, Preparation
of sputtered (Bax, Sr1x)TiO3 thin films directly on copper,
Journal of the American Ceramic Society, vol. 88.9.pp 26522654 2005.
[9] Malic, Barbara, Processing and dielectric characterization of
Ba0.3 Sr0.7 TiO3 thin films on alumina substrates, Journal of the
European Ceramic Society 27, pp. 2945-2948, 2007.
[10] Tahan, Danielle M., Ahmad Safari, and Lisa C. Klein.,
Preparation and characterization of BaxSr1x, TiO3 thin films
by a SolGel technique, Journal of the American Ceramic
Society, vol. 79.6, pp.1593-1598, 1996.
[11] Chen, Jun, Temperature dependence of piezoelectric properties
of high-TC Bi(Mg1/2Ti1/2)O3-PbTiO3, Journal of Applied
Physics, vol. 106.3 pp.034109-034109, 2009.
[12] IEEE Standard on piezoelectricity, IEEE Standard 176-1978
(IEEE, New York, 1978).
I would like to thank my father and teachers for their kind

and humble guidance from the very start to the end of the paper
281
Proceedings
A Comprehensive Study on QoS for Mobile

Ad Hoc Network
Abdur Rasheed
Irfan Nazeer
Fayyaz Khalid
Department of Computer Science

University of Gujrat
Gujrat, Pakistan
abdur.rasheed@live.com

Gujrat, Pakistan
irfan6200@yahoo.com

Gujrat, Pakistan
fayyazrao@yahoo.com
ABSTACT----A Mobile Ad Hoc Network (MANET) is a

collection of nodes which has no centralized
administration and without the assistant of base station.
All the nodes have equal rights and have their own
resources in MANET. QoS is a challenging issue now-adays due to bandwidth constraints and dynamic
topology in MANET. A lot of work done on supporting
on QoS in internet and other networks but most of them
is not suitable for MANET. Most of work is about the
QoS routing, QoS Mac and resource reservation, there
is a little work discuss on QoS models and its routing
protocols. In this article, we review the QoS models and
share some knowledge about QoS routing protocols and
other related work on QoS in MANET. The purpose of
this article is to give the appropriate knowledge about
QoS for MANET at the same platform.
mobile and nodes can move in any direction,

independent of each other. According to [3],
Mobile Ad hoc Networks are class of networks
whose contains characteristics is,
Physical characteristics As MANET is
a wireless network in which nodes can
create link in any time, bandwidth and
Delay due to the nodes are mobile.
Organization As mobile nodes are
distributed so resource estimation is not
predefined in
MANET
Dynamic Topology As the nodes have
dynamic topology so it has no fixed
architecture.
Keywords: QoS, Mobile Ad Hoc Networks
MANET is the combination of different mobiles

nodes which does not have any server, router or
access points for communication. It has low cast
wireless network and easy for deployment.
MANET is used in different situations in our life
i.e., Military operations, Rescue search
operations, communication between mobiles
vehicles in smart transportation, mobiles phones,
laptops and smart watches etc. As shown in Fig.1
I.
Introduction
Wireless network has become a popular network
now-a-days. Wireless communication has been
making our life easy and comfortable. Wireless
communication has been categorized into two
primary models; one is traditional wireless
network in which a fix infrastructure is used to
communicate between one hop to another. In
conventional wireless network, much of the
nodes are mobiles and connected with the help of
fixed
backbone
nodes
using
wireless
medium(radio waves). In this type wireless
network a base station is used for
communication. But in some situation (i.e. storm
disaster, battle field) there is no fix infrastructure
is used. Then we used a specific type of network
which is called Mobile Ad Hoc Network. A
mobile ad hoc network (MANET) [1][2], is a
dynamic wireless system in which all nodes are
MANET provides a flexible and seamless access to

internet. Due to in increasing popularity in MANET,
282
Proceedings
QoS has become the main task of research in

MANET. QoS is a challenging issue now-a-days due
to dynamic topology and bandwidth constraints. A lot
of work is done on supporting on QoS for internet
and other networks but most of them are not suitable
for MANET. Due to autonomous wireless network,
MANET has its own protocols.
circuit mechanism. Virtual connection is maintains

through signaling protocol RSVP (Resource
Reservation Protocol)[4].RSVP is very important
protocol in all QoS support routers. IntServ provides
two types of services; one is Guaranteed [5] Service
and the Other is Controlled Load Service [6]. Here
some pros and cons of IntServ mentioned below:
In this literature, we discuss QoS for MANET

includes QoS models and also discuss some
important protocols in MANET. There are many
researches on QoS on the internet. Due to this, it is
difficult to get appropriate knowledge on QoS in a
specific location. This paper provides a
comprehensive knowledge on QoS models for
MANET and some QoS routing protocols. In first
part, we shall discuss some QoS models which are
specified an architecture in which we discuss some
services that can be used in MANET. In second part,
we discuss some QoS protocols for MANET. These
protocol are divided in the based on route discovery,
metrics and network architecture. Then we gives a
comparison of these protocols on the bases on route
discovery, metrics and network architecture.
a)
Scalability
IntServ provides per-flow granularity so when state
information is increases the number of lows is also
increased. This requires storing a lot storage capacity
and processing is increases, it results a problem that's
called scalability problem of IntServ. The scalability
problem is very less in current MANETs because
there is less number of users, less bandwidth and
small size of network. On the other Hand, in future
the use of MANETs is increasing rapidly and the size
and speed of the MANETs increases so the
Scalability problem is also increased with time so
IntServ is not suitable in MANETs.
b)
Signaling
Signaling protocols like Resource Reservation
protocol have three steps; connection maintenance,
connection tear-down and connection establishment.
Corson[7] assumes and then suggest that a larger
number of links is given to the network to control
overhead .For MANETs with link capacities, the
overheads of connection maintenance and dynamic
topology usually a large for connection establishing
so that RSVP signaling protocol is not suitable in
MANETs.
II.
QoS models and MANETs
QoS model defines the architecture which provides a
platform to different services; which is used in
networks. QoS models use different situations like
dynamic topology etc. QoS model consider the
exiting internet architecture for MANET.
In this section, we explain the different QoS models
especially DiffServ and one of the newly proposed
QoS Model FQMM.
c)
Router mechanisms
IntServ is required a large number of routers so that
all router have packet scheduler, classifier, admission
control routine and RSVP that's why it is undesired
able in MANETs.
B. DiffServ and MANETs

Like ATM and IntServ, the tenet of DiffServ [8] is
used a relative-priority scheme to soften the hard
requirements of hard QoS models. DiffServ is
specially designed for the deficiencies of IntServ and
RSVP model like scalability problems in internet
backbone [9]. DiffServ is defined limited classes for
overcome this problem of IntServ. Much kind of
A. IntServ and MANETs

The idea of IntServ[4] is taken from the wired
network architecture and B-ISDN i.e. have the virtual
283
services are supported in DiffServ model which are

following [10]:
a) Premium Service
Premium Service is considered to give the end-to-end
bandwidth services, low jitter, low loss and low delay
[11].
c) Olympic Service
Three types of services are provided in Olympic
services which are Gold, Silver and Bronze with
decreasing quality [9].
Premium service is not suitable in MANETs because
it implementation in dynamic environment is not
possible. DiffServ is considered a defined solution to
MANETS QoS model because it provides Assured
Service and it is lightweight in inner router which is
suitable to MANETs. Because of this, DiffServ is
designed for wired networks and have some

problems in MANETs which are follows
Ingress Node
Interior Node
Egress Node
Ingress node is a source node which is a mobile node.

Interior node is the intermediate node that sends the
data of other mobile nodes. Egress node is the node
which is received data and it is the destination node.
Note that the role of Ingress node is changing due to
change in its position and network traffic. Resource
determination and allocation is done to the various
mobile nodes by using provisioning in FQMM.
Provisioning is at top level and based on utilization of
the network in present internet. Provisioning is used
to realize ideal per-flow by RSVP in IntServ and is
used in DiffServ as per-class by human agents rather
than RSVP or other. But on the other hand, FQMM
introduced a hybrid scheme as per-flow as well as per
class. Traffic with highest priority is given per-flow
provisioning while other priority is based on perclasses. Although it is the first efficient model in
MANETs but there is further need to study that how
sessions is entertained as per-flow due to bandwidth
constraints in MANETs.
b) Assured Service
Assured service is used to provide better reliability
for application. Other, it is more qualitative service
than quantitative and easy to implement.
Proceedings
III.
QoS Aware-Routing Protocols
In this section, we discuss some QoS aware routing
protocols. After this, we discuss their functionality
and then see their comparisons with each other with
different aspects. The QoS Aware-Routing Protocols
are follows:
DiffServ is not sure in to the boundary nodes for

this; it requires a lot of storage capacity.
DiffServ provides a Service Level Agreement
(SLA) which implementation is not possible in
MANETs. SLA is an agreement between
customer and the internet service Provider (ISP)
to receive Differentiated Services. SLA obeys
the partial or whole traffic rules [12]; but in
MANETs, it is not possible to negotiate the
traffic rules.
A. Optimized link stat routing Protocol

OLSR [14, 15] is a proactive routing protocols and
the idea of OLSR is given by IETF MANET[3]
working group. OLSR is the further study of Link
stat routing protocols. The idea of OLSR is based on
MPRs (Multi Point Relays).Every MPR have two
hops distance which is the part of communicating
nodes. If a node has no MPR then it cannot transfer
data further. Consider that MPR have larger distance
than a node which is not a MPR then we may miss
some good quality links and that's reason, it is
difficult to achieve a good quality of service in
OLSR.
C. Flexible Quality of Service Model (FQMM)

and MANETs
FQMM is the first QoS model which is preferred in
MANETs. FQMM is proposed in [13] and it tries to
take advantage from IntServ and DiffServ models
both. FQMM is designed for average size MANETs
which comprises less than 50 nodes using at nonhierarchical topology. However, FQMM is a hybrid
approach because it has combined features from
IntServ and DiffServ. In FQMM infrastructure, it has
types of Nodes like DiffServ which follows:
284
B. Predictive Location Based QoS Routing Protocol

PLBQR [16] is a proactive routing protocol which is
used in MANET.PLBQR is totally based on location
prediction of a node a particular instant in ad hoc
wireless networks. Location prediction is used to find
the geographical location of a node at particular
instant 'tf' when the receiver node picks a packet in
future. The value of 'tf' is calculated by calculating
delay in communication. No resources are reserved
during the propagation from source to destination but
QOS is controlled. This QoS routing protocol is
helped the nodes to update location prediction and
delay during communication.
Proceedings
route. If the route is deleted the new route request is

sent.
E. Adaptive QoS Routing Algorithm

ADQR [19] is a reactive protocol. Adaptive QoS
Routing algorithm protocol is used to establish a
route between two nodes which have longer life time.
In these protocols, the route breakage and route
creation is bases on signal power. Bandwidth is
considered to the lower layer. Route maintenance
contains two ways. One is pre-routing and the other
is rerouting. Pre-routing is used when the value of the
threshold is greater than signal power and rerouting is
used when the signal power is very low. For example
of this protocol is; if a node want to communicate
with other node then it sends a RREQ packet to
nearby node and then intermediate node attached his
address to this packet and forward to the next. If the
packet is reached to the exact destination then
destination node checked that the signal power.
Signal power ensures that the route is enough or it
disjoints with others .QoS_Reserve packet is sent to
the current route back and QoS Acknowledgment is
sent to the source. [3]
C. Ticket Based QoS Routing Protocol

Ticket Based QoS Routing protocol [17] is a reactive
protocol which is used in MANETs. The basic idea
of Ticket Based QoS Routing protocols is to use
tickets to find the best path which may delay
constrained or bandwidth constrained. This protocols
is used two types of tickets, one is yellow ticket and
the other is green. The purpose of yellow tickets is to
find the best path with delay/bandwidth constraints.
The other green ticket purpose is to find the low cost
route. The source node has a lot the number of tickets
to establish a QOS aware route. The tickets may be
less or more it depends on the route constraints to
find the best path. The Drawback of this routing
protocol is that the neighbor nodes incorporated to
each other with resource estimation thats why it
requires more memory. When a route is established
to the destination node acknowledged even resource
reservation is established.
F.
Trigger-based Distributed-QoS Routing

Protocol
TDR [20] is a reactive protocol which is a location
based protocol. In this protocol every host node
contains two types of databases. One is for local
neighbor and other is activity based. Hosts are passed
out the mobility and location information and the
neighbor nodes is stored the power level in their local
database. Activity-based data base is stored the every
session routing information and updated in every
session. Route discovery is done through the
neighboring power level and compared this with
threshold time. Route maintenance is done like
ADQR protocol [3].
D. Ad Hoc QoS on Demand Routing

The idea of AQOR [18] protocol is to find the best
route which is fulfills the constraints of end to end
delay with specific bandwidth. In route discovery, the
route request is sends to nearby node and attaches the
constraints with the packet and if both nodes satisfy
then this request is rebroadcasted to the next hop with
some expiry time. If the node is not received the
reply in given time then entry will be deleted. If the
nodes do not want to delete path the only
intermediate nodes will reply. Once the route is found
then resource reservation is takes place on the found
G. Bandwidth Estimation QoS Routing Protocol

BEQR [21] is a reactive protocol which is inspired by
AODV routing protocol [23].The main objective of
this protocol is to give soft QoS for service with the
bandwidth constraint. In this two schemes is used one
is Feedback scheme and the other is Admission
285
scheme. In Feedback scheme is used for updating

constraints when a node have no bandwidth its own.
Admission Scheme is used for route discovery with
satisfied bandwidth constraints. In this protocol
bandwidth is calculated with ratio of free and busy
times. Resource reservation is not considered. It is
design for soft QoS not hard.
sends the packets to the next node(intermediate).In

any time of transmission when topology is changed it
is established a new path. Every data packet has a
optimal QoS field and admission control module is
responsible for resource reservation; it is refreshed
periodically. If an application is used to want
minimum bandwidth then it bandwidth indicator is
"MIN", if the application is used to want maximum
bandwidth then indicator indicates "MAX".
H. Core Extraction Distributed Ad hoc Routing

Protocol
CEDAR [23, 3] was proposed by R. Siva Kumar
[23].CEDAR is a QoS aware routing protocol and it
is a hybrid routing protocol. In this protocol nodes
are selected as Core node and they are responsible for
route maintenance, computation and QoS providing
by using distributed algorithm. Link state information
depends on network state whether it is dynamic or
stable.
J. Forward Algorithm
FA [26] belongs to the on-demand routing protocols
which is used on demand algorithms with bandwidth
as QoS constraints parameter. FA is the modified
version of AODV [22] is used for route discovery
and some additional information is attached before
sending to the next hope. During route discovery, the
value of local maxima is calculated and forward [27].
FA is used in many other reactive protocols like
TORA [28] and DSR [29] for route discovery. The
Route maintenance is managed by using old routes in
it.
Route computation is on-demand based and Cedar

contain following three features which is very
important:
1.
2.
3.
Destination
location
discovery
and
establishing the core path to destination
Core nodes provides the stable and short
path between source to destination and also
a core path guideline
If the path is break then it re-established as
well as typology changes.
CEDAR considers that the resource is reserved.

[3]The route maintenance in CEDAR is done as
source initiated route maintenance or dynamic route
maintenance.
I.
Proceedings
INSIGNIA
INSIGIA [24] is a hybrid protocol which is proposed

for adaptive services in ad-hoc networks. Adaptive
services entertains that types of services which
requires minimum QoS guarantee (like minimum
bandwidth) and it varies when resources is sufficient.
INSIGNIA in-band signaling is the very important
component which is entertained fast reservation to
deliver the adaptive services[25].Routing module is
responsible for finding routes between nodes and
286
Proceedings
IV.
Comparison of QoS Aware-Routing Protocols
A comparison of QoS aware-routing protocols is given below:
Routing Protocols
QoS Support
Redundant Route
Route Maintenance
Route Breakage
Prediction
Network
Architecture
Route Discovery
OLSR
Hierarchical
Proactive
PLBQR
Proactive
Ticket Based
Location
Prediction
Flat
AQOR
ADQR
TDR
Flat
Flat
Location Based
Reactive
Reactive
Reactive
BEQR
Flat
Reactive
CEDAR
Hierarchical
Hybrid
INGIGNIA
Flat
Reactive
FA
Flat
Reactive
Conclusion
In first part of this paper, we have described three
QoS Models (IntServ, DiffServ and FQMM).All
three models are suitable for different situations. All
three have some pros and cons; but FQMM is a
flexible QoS model for MANETs. To our knowledge,
FQMM was the first such QoS model proposed for
MANETs. It provides a hybrid provisioning scheme
and a relative and adaptive traffic profile. In second
part of this paper, we described some important QoS
aware-routing protocols in ad hoc wireless networks
which is an active research now-a-days. In this paper,
we discussed some characteristics of the QoS routing
protocols and compare the characteristics of these
protocols in the form of table. The aim of this review
is to give an appropriate knowledge on QoS for
MANETs at the single paper.
Reactive
[5] S. Shenker, C. Partridge and R. Guerin Specification of

Guaranteed Quality of service RFC 2212 September 1997
[6] J. Wroclawski, Specification of the Controlled-Load Network
Element Sevice RFC 2211, Sept.1997
[7] M.S. Corson. Issues in supporting quality of service in mobile
ad hoc networks. In IFIP 5th Int.Workshop on Quality of Service
(IWQOS'97),May 1997.
[8] S. Blake. An architecture for Di_erentiated Services. IETF
RFC2475, December 1998.
[9] Xipeng Xiao Nad Lionel M. Ni Internet QoS: a big Picture,
IEEE Network magazine , March 1999.
[10] Kui Wu and Janelle Harms Quality of Services in Mobile Ad
Hoc Networks
[11] K. Nichols, V. Jacobson, and L. Zhang. A two-bit
Di_erentiated Services architecture for the Internet. IETF
RFC2638, July 1999
[12] S. Blake, An Architecture for Differentiated Services, IETF
RFC 2475, December 1998.
References
[13] H. Xiao , W.K.G. Seah, A. Lo and K. C. Chua, A flexible

Quality of Service Model for Mobile Ad hoc Networks IEEE
VTC2000-spring . Tokyo, Japan May 2000.
[1] Loay Abusalah, Ashfaq Khokhar, and Mohsen Guizani A

Survey of Secure Mobile Ad Hoc Routing Protocols IEEE
communications surveys & tutorials, VOL. 10, NO. 4, FOURTH
QUARTER, 2008, pp 78-93.
[14] Y. Ge, T. Kunz, and L. Lamont, Quality of Service Routing

in Ad-Hoc Networks Using OLSR, Proc. 36th Hawaii Intl. Conf.
Sys. Sci., Jan. 2003.
[2] Sunil Taneja and Ashwani Kush A Survey of Routing

Protocols in Mobile Ad Hoc Networks, International Journal of
Innovation, Management and Technology, Vol. 1, No. 3, August
2010.
[15] P. Jacquet, P. Muhlethaler, A. Qayyum, A. Laouiti, L.

Viennot, T. Clauseen, "Optimized Link State Routing Protocol
draft-ietf-manet-olsr-05.txt", INTERNET-DRAFT, IETF MANET
Working Group.
[3] Lei Chen and Wendi B. Heinzelman, A Survey of Routing

Protocols that Support QoS in Mobile Ad Hoc Networks, IEEE
Network November/December 2007.
[4] R. Braden, D. Clark, and S. Shenker. Integrated Services in the
Internet architecture - an
Overview. IETF RFC1633, June 1994.
[16] S.H.Shah and K.Nahrstedt, Predictive Location- Based QoS

Routing in Mobile Ad Hoc Networks, Proceedings of IEEE ICC
2002,vol.2, pp.1022-1027, May 2002
287
[17] S. Chen and K. Nahrstedt, Distributed Quality-of-Service in

Ad Hoc Networks, IEEE JSAC, vol. 17, no. 8, 1999.
[18] Q. Xue and A. Ganz, Ad hoc QoS On-demand Routing
(AQOR) in Mobile Ad hoc Networks, J. Parallel and Distrib.
Comp., vol. 62,no. 2, Feb.2003, pp. 15465.
[19] Y. Hwang and P. Varshney, An Adaptive QoS Routing
Protocol with Dispersity for Ad-hoc Networks, Proc. 36th Hawaii
Intl.Conf. Sys. Sci., Jan. 2003.
[20] S. De et al., Trigger-Based Distributed QoS Routing in
Mobile Ad Hoc Networks, ACM SIGMOBILE Mobile Comp.
and Commun.Rev., vol. 6, no. 3, July 2002, pp. 2235.
[21] Lei Chen and Wendi B. Heinzelman, QoS-Aware Routing
Based on Bandwidth Estimation for Mobile Ad Hoc Networks,.
IEEE Journal on Selected Areas in Communications,
23(3):561_572, March 2005.
[22] C. E. Perkins and E. M. Royer. Ad-hoc On-Demand Distance
Vector Routing, In Proceedings of the 2nd IEEE Workshop on
Mobile Computing Systems and Applications, pages 90_100, New
Orleans, LA, Feb. 1999
[23] R. Sivakumar, P. Sinha, and V. Bharghavan, CEDAR: a core
extraction distributed ad hoc routing algorithm IEEE Journal on
Selected Areas in Communication, Vol. 17, no. 8, pp 1454-65
August1999.
[24] X. Zhang, S.B Le, A. Gahng-Seop and A.T Campbell,
INSIGNIA: an IP based quality of service framework for mobile
ad hoc networks.Journal of Parallel and Distributed Computing,
Vol. 60, no. 4,pp. 374-406, April 2000.
[25] C.Siva Ram Murthy,B.S.Manoj, Ad Hoc Wireless Networks
Architectures and Protocols Prentice Hall Communications
Engineering and Emerging Technologies Series.
[26] Chenxi Zhu and M. Scott Corson. QoS routing for mobile ad
hoc networks. Technical report, Flarion Technologies,
Bedminster, New Jersey 07921, 2002.
[27] Philipp Becker, QoS Routing Protocols for Mobile Ad-hoc
Networks A Survey, TU KL Technical Report 368/08, A
publication by University of Kaiserslautern
[28] V.D. Park and M.S. Corson, A Highly Adaptive Distributed
Routing Algorithm for Mobile Wireless Networks, Proceedings of
IEEE INFOCOM 1997, pp.1405-1413, April 1997.
[29] D.B.Johnson and D.A. Maltz, Dynamic Source Routing in
Ad Hoc Wireless Networks, Mobile Computing, Kluwer
Academic Publishers,vol.353,spp.153-181,1996.
288
Proceedings
Proceedings
Comparison of Maximum Likelihood Classification

Before and After Applying Weierstrass Transform
Muhammad Shoaib, Zaka Ur Rehman, Muhammad Qasim
Abstract The aim of this paper is to use Maximum
Likelihood (ML) Classification on multispectral data by means of
qualitative and quantitative approaches. Maximum Likelihood is
a supervised classification algorithm which is based on the
Classical Bayes theorem. It makes use of a discriminant function
to assign pixel to the class with the highest likelihood. Class
means vector and covariance matrix are the key inputs to the
function and can be estimated from training pixels of a particular
class. As Maximum Likelihood need some assumptions before it
has to be applied on the data. In this paper we will compare the
results of Maximum Likelihood Classification (ML) before apply
the Weierstrass Transform and apply Weierstrass Transform
and will see the difference between the accuracy on training
pixels of high resolution Quickbird satellite image. Principle
Component analysis (PCA) is also used for dimension reduction
and also used to check the variation in bands. The results shows
that the separation between mean of the classes in the decision
space is to be the main factor that leads to the high classification
accuracy of Maximum Likelihood (ML) after using Weierstrass
Transform than without using it.
Index Terms Maximum Likelihood Classificaion (ML),
Weierstrass
Transform,
Bayes
Theorem,
Supervised
Classification, Principle Component Analysis (PCA).
I. INTRODUCTION
Satellite images are widely used for various purposes and
with the advancement in Satellite technology every country
wants to solve their problem by using satellite images. Satellite
images can be used in various disciplines such as Metrology,
Environmental sciences, GIS, Geology, Geophysics,
Engineering and Construction, Defense and Intelligence and
the list goes on.
Satellite images contains maximum information about our
planet and main problem is that how can we retrieve the useful
information satellite images. Researchers can have several
problems such how to classify the land cover types, depth of
Aerosol particles in atmosphere, Forest Climates, Flood
predictions and many more. Researcher solve these problems
by analyzing the satellite images.
In this article our main aim is to classify the land cover
types by using Maximum Likelihood Classification algorithm.
First we classify land cover types without using Weierstrass
Transform (WT) and then we compare the results by applying
Weierstrass Transform (WT).
Maximum Likelihood (ML) is a supervised classification
method derived from the Bayes theorem, which states that the a
posteriori distribution P(i|), i.e., the probability that a pixel
with feature vector belongs to class i, is given by:
where P(|i), is the likelihood function, P(i) is the a priori

information, i.e., the probability that class i occurs in the study
area and P() is the probability that is observed, which can be
written as:
where M is the number of classes. P() is often treated as a

normalization constant to ensure P(i|) sums to 1. Pixel x is
assigned to class i by the rule:
Each pixel is assigned to the class with the highest likelihood

or labeled as unclassified if the probability values are all
below a threshold set by the user [1].
Maximum Likelihood (ML) assumes that the data within a
given class i obeys a multivariate Gaussian distribution.
Frist we implement Maximum Likelihood (ML) classification
algorithm without checking the assumption of normality and
then we apply Maximum Likelihood (ML) classification after
using Weierstrass Transform on the image data and we have
seen that after applying Weierstrass Transform our data is
normally distributed. The accuracy after applying Weierstrass
Transform (WT) is increased because the distribution of the
data is normal because for ML classification data has to be
normal.
II. STUDY AREA
We use Quickbird satellite image for our analysis.
Quickbird is high-resolution commercial satellite which is
owned by DigitalGlobe and launched in 2001. Quickbird uses
Aerospace's Global Imaging System 2000 (BGIS 2000).The
Quickbird satellite collected panchromatic (black and white) at
61cm resolutin and multispectral imagery at 2.44-(at 450 km)
to 1.63-meter (at 300km) resolution, as orbit altitude is lowered
during the end of mission life [2].
The Quickbird satellite image which we use for our
analysis is sample image of Rome Italy with coordinates
with four bands which are
blue (450-520nm), green (520-600nm), red (630-690 nm) and
NIR (760-900 nm).
289
Proceedings
III. METHODOLOGY
Blue Band Histogram
Methodology is the systematic, theoretical analysis of the

methods applied to a field of study. It comprises the theoretical
analysis of the body of methods and principles associated with
a branch of knowledge (wiki).
The detail methodology is given in the figure 1.
14000
Blue Band
12000
10000
8000
6000
4000
Distribution Check
Preprocessing
2000
0
0
Weierstrass Transform
50
100
150
200
250
Green Band Histogram

14000
12000
Training Samples
PCA
Green Band
10000
8000
6000
Results
4000
2000
ML with WT
ML without WT
100
200
Red Band Histogram

14000
Conclusions
12000
Red Band
Fig. 1. Data Processing flowchart
10000
8000
IV. PREPROCESSING
6000
Preprocessing is a common name for operations with

images at lowest level of abstraction both input and output are
intensity images. The main aim of preprocessing is an
improvement of the image data that suppresses unwanted
distortions or enhances some image features important for
further processing.
In this paper we have to do classification so we do not need
to do any radiometric calibration or anything else. But we have
to check the distribution of each band whether it is normal or
not which is the main assumption for ML classification.
4000
2000
0
0
100
200
NIR Band Histogram

12000
10000
A. Distribution Check
We use image histogram to check the distribution of each
band. An "image histogram" is a type of histogram that acts as
a graphical representation of the tonal distribution in a digital
image. It plots the number of pixels for each tonal value. By
looking at the histogram for a specific image a viewer will be
able to judge the entire tonal distribution at a glance. We make
a separate histogram for each band to check the distribution of
each band in image.
NIR Band
8000
6000
4000
2000
0
0
100
Fig. 2. Histograms with their respective bands
290
200
In the figure 2 we can see that the distribution of each band

is not Gaussian and we know that for ML classification
algorithm the distribution of each has to be Gaussian. In this
case distribution of first three bands are positively skewed
while the histogram of band 4 is looks like Gaussian but not
exactly Gaussian. So as our aim is to check the accuracy of
classification before making data Gaussian and after making it
Gaussian. We use Weierstrass Transform to make the
distribution Gaussian.
Proceedings
combinations of X and Y are normal in order to conclude that

the vector [X Y] is bivariate normal.
Gaussian filtering is used to blur images and remove noise
and detail. Gaussian filter uses the above Gaussian function
which is given in equation 1.
After applying Gauss transform or Wierstrass transform we
can compare original false color composite (FCC) image with
transformed false color composite (FCC) image.
B. Weirstrass Transform
As we have seen that distribution of our data is not follow
Gaussian distribution so make that Gaussian we have to use
transformation to make pixel values Gaussian.
Researchers uses various types of transformation to make
the data Gaussian. In this paper we use different type of
transformation called Weirstrass Transform.
The Weirstrass transformation, also known as the Gauss
transform, the Gauss Weirstrass transform and the Hille
transform, is intimately related to the solution of the heat
equation for one- dimensional flow. It is also a special case of
convolution-transform, yet it is considered a singular case of
convolution-transform theory developed by Hirschmann and
Widder [3].
The Weirstrass transformation F(x) of a function f(y)
defined as:
(a) original false color composite (FCC) image
the convolution of f with the Gaussian function
Instead of F(x) we also write W[f](x). Note that F(x) need not
exist for every real number x, because the defining integral
may fail to converge [3].
When Gaussian filter modifies the input signal by using
Gaussian function; this is known as Weierstrass
transformation.
We can write Gaussian function mathematically in 2D as
follow:
(b) Transformed false color composite (FCC) image

where is the correlation between X and Y and where _x>0
and _y>0. In this case,
Fig. 3.
In figure 3 we can see the clear difference between original

FCC image with transformed FCC image. After applying
transformation we can see that the transformed image is more
blur than original image.
In the bivariate case, the first equivalent condition for

multivariate normality can be made less restrictive: it is
sufficient to verify that countably many distinct linear
291
Proceedings
4
x 10 Red Band Histogram
4
x 10 Blue Band Histogram
2
1.8
1.6
Red Band
Blue Band
1.5
1.4
1.2
1
0.8
0.6
0.5
0.4
0.2
0
0
0
0.5
0.5
(c)
(a)
4
x 10 Green Band Histogram

NIR Band Histogram
2
18000
1.8
16000
Green Band
1.6
NIR Band
14000
1.4
12000
1.2
10000
1
8000
0.8
6000
0.6
4000
0.4
2000
0.2
0
0
0.5
0.5
(d)
(b)
Fig. 4. (a) Histogram of band 1 of transformed image, (b) Histogram of band 2 of transformed image, (c) Histogram of band 3 of
transformed image, (d) Histogram of band 4 of transformed image
In Figure 4 if we look at the histograms of different bands

of transformed false color composite (FCC) image we can
examine that the distribution of each band follow Gaussian
distribution. We can see now pixel values of image follow
standard Gaussian distribution.
C. Dimension Reduction
Dimension reduction in satellite image processing reducing
the number of bands which have same information in nature.
Dimension Reduction is used when we have large number of
bands.
We have Quickbird image which have 4 bands and we
apply PCA to check which band we have to ignore. We apply
PCA on transformed FCC image.
Fig. 4. Principle Components of four bands of transformed
image.
292
In the above figure we can see that band 3(Red) and band
4(NIR) have same information so we have to neglect band 3
or band 4 from our analysis because they have same
information in the image for this purpose we will use
classification on three bands that are Blue, Green and NIR.
D. Training Samples
Training areas were established by choosing one or more
polygons for each class. Pixels fall within the training area
were taken to be the training pixels for a particular class. In
order to select a good training area for a class, the important
properties taken into consideration are its uniformity and how
well they represent the same class throughout the whole image
[4].
We use Stratified random sampling for selecting the
training pixels. A stratified random sample is a population
sample that requires the population to be divided into smaller
groups, called 'strata'. Random samples can be taken from each
stratum, or group.
Proceedings
parallelepiped classification algorithm. But as we know that

ML classification algorithm assumes that distribution of each
band has to be Gaussian. we had seen that our band's
distribution was not Gaussian so we make distribution
Gaussian by applying Gauss transform or Weirstrass transform.
Now we will check what is the effect on accuracy of
classification classes when the distribution is Gaussian and
when not Gaussian.
A. Classification Before Weirstrass Transform
The outcome of ML classification after assigning the
classes with suitable colors, is shown in Fig.5 (a): Grass
(green), Urban Area (blue), Roads(yellow), Soil (cyan), Trees
(red) and Unclassified (Black). The areas in terms of
percentage and square meters were also computed; the classes
with the largest area is Urban Area. In this case the distribution
of our data is not Gaussian and we can clearly see in the Fig.5.
(a) that some buildings are wrongly classified as Roads
because we are applying ML classification on non normal data.
V. RESULTS
Maximum Likelihood considered to be an most accurate
algorithm when we compare it to classical algorithm such as
Class
Unclassified:
Trees
Grass
Urban Area
Roads
Soil
Color
Area in %
[Black]
6.44%
[Red]
14.20%
[Green]
13.63%
[Blue]
31.45%
[Yellow]
30.31%
[Cyan]
3.97%
Area in Meters)
Color
(26,783.6400 Meters)
(59,095.0800 Meters)
(56,719.0800 Meters)
(130,868.2800 Meters)
(126,133.9200 Meters)
(16,532.6400 Meters)
Color
Area in %
[Black]
10.30%
[Red]
13.75%
[Green]
10.34%
[Blue]
44.79%
[Yellow]
19.39%
[Cyan]
1.44%
Area in Meters)
Color
(42,851.1600 Meters)
(57,232.0800 Meters)
(43,005.2400 Meters)
(186,373.0800 Meters)
(80,677.4400 Meters)
(5,993.6400 Meters)
(a)
Class
Unclassified:
Trees
Grass
Urban Area
Roads
Soil
(b)
Fig. 5. (a) Maximum Likelihood classification an original image, (b) Maximum Likelihood classification on transformed image
293
TABLE I.
Proceedings
CONFUSION MATRIX OF ORIGINAL CLASSIFIED IMAGE
Overall Accuracy = (18318/21394) = 85.62%

Kappa Coefficient = 0.8135
Ground Truth (Percent)
Class
Color
Trees
Grass
Urban Area Roads
Soil
Color
Unclassified [Black]
0.4
0.05
7.84
2.33
0
Trees
[Red]
94.38
8.06
0.94
5.28
0
Grass
[Green]
1.64
88.71
2.66
0.03
3.41
Urban Area [Blue]
0
1.06
70.71
1.19
2
Roads
[Yellow
3.58
0
4.61
91.17
0
Soil
[Cyan]
0
2.11
13.23
0
94.59
Total
100
100
100
100
100
TABLE II. CONFUSION MATRIX OF TRANSFORMED CLASSIFIED IMAGE
Overall Accuracy = (20156/21394) = 94.2133%

Kappa Coefficient = 0.9236
Ground Truth (Percent)
Class
Color
Trees
Grass
Urban Area Roads
Soil
Color
Unclassified [Black]
0.85
0.36
4.36
0.52
0
Trees
[Red]
98.51
4.72
0.09
0.08
0
Grass
[Green]
0.65
94.71
1.92
0
0.3
Urban Area [Blue]
0
0.2
86.5
1.33
0.52
Roads
[Yellow]
0
0
3.19
98.07
0
Soil
[Cyan]
0
0
3.94
0
99.19
Total
100
100
100
100
100
Accuracy assessment of the ML classification was
determined by means of a confusion matrix (sometimes called
error matrix), which compares, on a class-by class basis, the
relationship between reference data (ground truth) and the
corresponding results of a classification [1].
The results are given in Table 1. The diagonal elements in
Table 1 represent the percentage of correctly assigned and are
also known as the producer accuracy. Producer accuracy is a
measure of the accuracy of a particular classification scheme
and shows the percentage of a particular ground class that is
correctly classified.
The Kappa coefficient, is a second measure of
classification accuracy which incorporates the off-diagonal
elements as well as the diagonal terms to give a more robust
assessment of accuracy than overall accuracy.
The ML classification yielded an overall accuracy of
85.62% and kappa coefficient 0.8135.
B. Classification After Weirstrass Transform
When we apply Gauss transform on data we have already
seen that our band's distribution changes from skew symmetric
to approximately Gaussian.
When we apply ML classification algorithm on

transformed image we have seen that the overall accuracy
increases from 85.62% to 94.21% because pixel values of
image data follows Gaussian distribution and also the value of
Kappa coefficient, increases from 0.8135 to 0.9236. In Table
2 we can see the results of classification after applying the
Weirstrass Transform or Gauss transform. So when our data
follows Gaussian distribution we can say that ML classification
produce better.
VI. CONCLUSIONS
In this study, detail analyses of ML classification for
tropical land covers in Rome have been carried out, in which
lead to a number of conclusions. ML classifies the classes that
exist in the study area with a good agreement when distribution
follows Gaussian and when distributions is not Gaussian we
can see that overall accuracy decreases almost 9%. ML
classified the study area into 5 classes, with accuracy 88% (=
0.8425). ML classifies pixels based on known properties of
each cover type, but the generated classes may not be
statistically separable.
294
ACKNOWLEDGMENT
The authors would like to thanks the Dr Waqas A. Qazi for
his valuable input and ideas, Muhammad Qasim for helping
with the implementation of Weirstrass transform.
REFERENCES
[1] T.M. Lillesand, R.W. Kiefer and J.W. Chipman, Remote
Sensing and Image Interpretation, John Wiley & Sons,
Hoboken, NJ, USA, 2004.
[2] DigitalGlobe.com. DigitalGlobe. 12 February 2014. Retrieved
19 June 2014.
[3] Ahmed I. Zayed, Handbook of Function and Generalized
Function Transformations, Chapter 18. p.322-324.
[4] J.B. Campbell, Introduction to remote sensing, Taylor &
Francis, London, 2002.
Muhammad Shoaib is Master student of RS and GIS in

the department of Space Science in Institute of Space
Technology Islamabad, Pakistan.
Zaka Ur Rehman is Master student of RS and GIS in the
department of Space Science in Institute of Space Technology
Islamabad, Pakistan.
Muhammad Qasim received the Masters degree in
Mathematics from Comsats Institute of information
technology Abbottabad, Pakistan.
295
Proceedings
Proceedings
Spatio -Temporal Analysis of Shoreline Changes

along Makran Coast Using Remote Sensing and
Geographical Information System
Hira Fatima1, Mudassar H. Arsalan, Anam Khalid, Kashif Marjan, Mukesh Kumar
Geo-informatics and Sustainable Development Research Lab
Institute of Space and Planetary Astrophysics
University of Karachi
Karachi, Pakistan
1
hirafatima1407@gmail.com
AbstractSatellite Images of Makran coast captured by

Landsat 5 TM in 1987, 2000 and 2010 were integrated in a GIS
(Geographical Information System) to perform spatio-temporal
analysis of shoreline changes, coastal accretion and erosion, and
coastline length. It was found that Remote sensing data and GIS
are time-saving and cost effective for such analyses. Makran
shoreline has changed noticeably from 1987 to 2010. During
1987-2010, erosion was dominant along western Makran, it was
eroded at the rate of 0.2 sq. km per year. However accretion was
dominant along Eastern Makran coast, it was advanced at the
rate of 0.5 sq. km per year. During the long time span of 23 years
(1987-2010) Gwadar port was eroded 4 sq. km which was quite
higher than that of Jiwani i.e. 2.4 sq. km. However Ormara,
Pasni and Astola Island were relatively stable places.
Index Terms Shoreline change detection, Erosion - accretion
analysis, Image interpretation, Band ratio index, Satellite images,
Geospatial technologies, GIS, RS, Geo-informatics.
I. INTRODUCTION
A shoreline being the interface between land and water is a
dynamic feature on the earth's surface. It is always changing
due to sediments continually being eroded from one place and
deposited to another. Many factors play a part behind these
changes, they may be natural like waves, winds, tides and
storms as well as human influenced such as dredging, mineral
exploration, vegetation removal, infrastructure development
etc. These natural processes and human activities can change
shoreline on small time scales i.e. days and years or long time
scales such as decades or centuries. Monitoring shoreline
changes, is very crucial in coastal management and planning.
The rate of these changes i.e. erosion and accretion, can be
used to forecast future shoreline trends. Decision makers need
this information to assess the feasibility of a project and
Government organizations for disaster risk management [1-5].
Remote sensing and Geographical Information System are

widely used for shoreline change detection analysis. The
spectral, spatial and temporal resolutions, repetitive coverage,
synoptic view and cost effectiveness of remote sensing data has
made it more favorable than data collected by conventional
techniques. The shoreline change detection analysis is now
widely being done using diversified data such as historical
topographical maps, aerial photographs, Digital Elevation
Models (DEM), and satellite images. Satellite images from
different satellites (such as Landsat 5, Landsat 7, Landsat 8,
IKONOS, Quick bird, IRS etc.) have been used for such
analysis [5-11]. Further, the data management, visualization
and data analysis capabilities of GIS are applied on the data
derived from satellite images to get desirable results [12-14].
The initial and crucial step in the shoreline change
detection is the selection of appropriate feature that can be
treated as shoreline proxy or indicator that can represent the
current position and can act as a reference to monitor the
change in different periods [2, 15]. These proxy shorelines are
typically of two types: a visible and distinguishable line in
coastal imagery known as High Water Line (HWL) or
intersection of a tidal datum with coastal profile (say MHW,
MLW, etc.) [2]. Among all the vertical levels that are used in
the shoreline change detection, HWL is very common as it is
easily interpretable in photos [2]. After identification, shoreline
can be extracted by manual digitization or some automated
techniques in GIS [13, 16]. Further, the shoreline movement,
erosion and accretion rates can be calculated and its future
trends can be predicted [17, 18].
Literature review shows shoreline change detection is now
considered as an important part of sustainable coastal resource
management and environment conservation, hence it is
monitored throughout the World [19]. Table 1 gives a
summary of several shore detection studies done throughout
the world for different parts of coasts such as beaches, river
296
mouths, deltas, lakes, estuaries and even islands, and geospatial

data and techniques used in these case studies.
Over the 100 years, about 70% of worlds sandy coasts
(which constitute about 20% of the worlds coastline) have
been retreating while 20-30% have been stable and less the
10% have been advancing [20]. Coastal erosion is notable
along the sandy beaches of Pakistan [21]. Literature review
shows that very few shoreline change detection studies for
coastal areas of Pakistan have been done [22, 23] and there is a
big research gap in this dimension. This paper aims to bridge
this research gap.
The Pakistan coast consists of Makran coast (the major
portion) and the Sindh coast. Sindh coast is further divided into
Indus deltaic coastline (a network of 17 minor creeks) and the
Karachi coastline that extends from West of Indus delta to the
outfall of the Hub River from where the Makran coast starts
and extends up to the Iranian border [21, 24-26].
The coastal length of Pakistan found in literature is
ambiguous. It varies from 825 to 1200 km in different sources
(see Table 2). Literature review shows that reason behind this
ambiguity is coastline paradox. This phenomena was first
discovered by Lewis Fry Richardson [27]. Coastline does not
have a well-defined length because of its fractal like
characteristics. The coastline has features at very different
scales, hence the length of coastline increases every time when
it is measured at a smaller scale [28]. In short, the length of
coastline depends on the method used to measure it, the scale
and the resolution of data used [28]. The length of coastline
also varies as the shoreline evolves. with some places say
beaches getting sediments to deposit while others getting
eroded [29]. However, with the advent of geospatial
technologies and high resolution remote sensing data worlds
coastlines can be measured more precisely and this paradox
can be handled to some extent. Not only the length of coastline,
but the area of sea can also be monitored efficiently and cost
effectively using remote sensing data and GIS [30]. In this
study, the length of Makran coast, the study area, in different
years (i.e. 1987, 2000 and 2010) were calculated from Landsat
TM satellite images. However, more precise coastline length
measurements can be made using high resolution data, such as
SPOT [31].
Proceedings
A. Study Area
The coastal area of Pakistan selected for this study is
Makran coast (see Figure 2 and 6). Most of the Makran coast is
undeveloped, with deserted beaches and some fishing villages.
Further, it is characterized by unique landforms such as mud
flats, rocky cliffs, bays, lagoons, deltas, as well as of the
narrow strip of mountains, which is known as Makran ranges.
Makran ranges cover area of about 400 km long and 250 km
wide. These mountains have elevation up to 1500 m above
mean sea level [26, 32]. The Makran coast is blessed with
scenic beauty of several beaches: Somiani, Hingol River,
Ormara, Pasni, Gwadar etc. The main towns and fishing ports
of the study area are Ormara, Pasni and Jiwani. Gwadar deep
sea port is the economic region of Balochistan coast. It is a
major destination in the Pakistan-China economic corridor [32,
33].
Makran coast differs Sindh coast in several ways. Indus is
the only major river of Pakistan and its delta lie at the Sindh
coast, while at Makran coast only small rivers i.e. Hingol, Hab,
Basul and Dasht do exist. The depths along the Sindh coast
change moderately, while the Makran coast is steep. The
average rainfall on the Makran coast is about 10 mm while it is
twice at the Sindh coast [34]. Makran coast is less vulnerable to
sea level rise than Sindh coastal area as the Makran coast is
uplifting about 1-2 mm/year due to the subduction of Indian
oceanic plate [22]. Coastal erosion is one of the typical effect
of sea level rise [35]. A very threatening subduction zone is
located about 100 km away from Makran coast that can cause
tsunami [36, 37]. In 1945 tsunami, Las Bela State of
Balochistan (which is now Western Pakistan) was affected
badly, particularly the Pasni and Ormara towns [38].
Pakistan has mixed semi-diurnal tides, however the tidal
range varies throughout the coastline [22]. The tidal amplitudes
are greater in the Indus deltaic region as the sea water flows
into creeks with high velocity during flood and ebb tides,
however the tidal amplitudes gradually decreases towards west
(Makran coast) [22, 39]. During the Southwest monsoon
season the direction of the prevailing ocean current is
clockwise and counter clockwise during the Northeast
monsoon season in the Arabian Sea [34]. Generally, the
salinity value is approximately 36 ppt [36].
In this study, we examined different bands of Landsat 5
TM, their combination and ratios for shoreline delineation. We
calculated shoreline length for different years and studied
coastline paradox phenomenon. We analyzed the long and
short term shoreline changes and associated erosion and
accretion along the Makran coast using Landsat TM data
acquired in 1987, 2000 and 2010.
Fig. 1. Coastline Paradox Illustration
In this study, we used Landsat 5 TM data and two commercial

software (i.e. ArcGIS 10.3 and ERDAS Imagine 9.2) for data
management, analysis and visualization. The objectives of
this study were achieved by simple steps shown in the figure
3.
297
Proceedings
Table 1 Shoreline Change Detection Studies throughout the World

Serial
#
Author
Year
Study Area
Technique used
Data Used
Shetty
[40]
2015
Mangalore Coast,
Netravathi-Gurupur and
Mulky-pavanje Spits
Visual interpretation and on screen

digitization
Landsat 5, IRS and Landsat 8
Natesan
[41]
2015
Tamil Nadu, India
Layer Stacking, digitization
Hegde
[42]
2015
Karnataka Coast, India
Aedla
[16]
2015
Netravati-Gurpur River
mouth
Mahapat
[10]
Choudharey
[43]
Murray
[8]
2013
South Gujarat Coast
2013
from Karwar to Gokarna South West coast of India

Wotje Atoll, Marshall
Islands
Histogram slicing of infrared band,

unsupervised classification and
digitization
Modified SelfAdaptive Plateau
Histogram Equalization with Mean
Threshold
False color composite (bands 2, 3, 4),
visual interpretation and digitization
Geo referencing, digitization
Landsat MSS, Landsat TM,

Landsat ETM+ and Landsat
OLI
Landsat TM, Landsat ETM+,
Landsat OLI-TIRS
Rio
[4]
Das
[44]
2013
SW Spain
2013
Shankarpur - Mandarmoni
Coast line, West Bengal
Azaz
[45]
Faiboon
[46]
2012
Wedam_Alsahel area,
Batinah, Oman
Chalatat Beach in
Songkhla Province,
Thailand
Eludoyin
[1]
Kumaravel
[47]
Kuleli
[6]
2012
15
Adegoke
[3]
2010
Niger Delta, Nigeria
16
Li
[48]
2010
Pearl River Estuary, China
17
Maiti
[5]
2009
Bay of Bengal, Eastern

India
18
Kumar
[49]
Rajamanickkam
2009
Mangalore Coast, India
2006
Klien
[50]
Makota
[13]
Frazier
[51]
2006
2004
Kallar and Vaipar Coast,

Tamilnadu, India,
Position adjacent to Hadera
power station, Israel
Kunduchi Area, Tanzania
2000
Channel and flood plain of

Murrumbidgee River
6
7
8
9
10
11
12
13
14
19
20
21
22
2013
2012
2012
2011
Bonny Island, Nigeria

Cuddalore district, East
coast of Tamil Nadu, India
Coastal Ramsar Wetlands
of Turkey
Mosaicking, Digitization
Georeferencing, polynomial
correction and digitization
Band ratio mode between mid
infrared and green, digitization
Geometric correction, Image
degradation, digitization
Geometric correction, resampling,
Overlay
Layer stacking, visual interpretation
and digitization
Georeferencing, layer stacking,
digitization
Top of atmosphere segmentation
algorithm, NDWI, Automatic
thresholding and shoreline extraction
Radiometric and geometric
correction, georeferencing,
mosaicing, Histogram equalization,
Band combination 6,4,2 and
digitization
Georeferencing, Mosaicking, Layer
stacking, slicing, classification and
digitization
Georeferencing, gray level
thresholding and segmentation by the
edge enhancement technique of NIR
bands, Digitization
Georeferencing, layer stacking,
digitization
Edge enhancement, level slicing,
NDVI, digitization
Rectification and georeferencing,
Digitization
Visual interpretation and on screen
digitization
Single band density slicing, multispectral maximum likelihood
algorithm, digitization
298
IRS data

Landsat ETM+, IRS
Toposheet and IRS data
Aerial photographs,
IKONOS, QuickBird,
GeoEye-1
Aerial Photographs
Topomap, Landsat TM,
Landsat ETM+ and Google
Earth imagery
IRS, IKONOS
Landsat ETM+, Landsat TM,
Landsat TM, Nigersat data

Toposheets, IRS data
Landsat ETM+
Landsat TM, Landsat ETM+
data
Topographic map, nautical

map, Landsat TM, Landsat
MSS, Landsat TM+, Spot XS
Toposheets, Landsat MSS,
Landsat TM, Landsat ETM+
and ASTER
Toposheets, IRS data
Topo sheets and IRS data
Aerial Photographs
Aerial Photographs
Landsat 5 TM data, Aerial
Photographs
42, 154-43, 153- 42 and 153- 43) acquired in 1987, 2000 and
2010. Landsat 5 was a joint effort of NASA and USGS,
launched on 1st March 1984. Landsat 5 was put in a circular,
sun synchronous, near polar orbit at the altitude of 705.3 km.
Landsat 5 carried TM (Thematic Mapper) and MSS (Multi
Spectral Scanner System) sensors. It took approximately 16
days to scan the whole earth's surface. It used to cross equator
at 9:45 a.m. (+/- 15 min). [57]
Landsat 5 was decommissioned on 5th June 2013, due to
several mechanical failures. Landsat 5, in its 29 year, 3 months,
4 day journey orbited 150,000 times, transmitted over 2.5
million images. Several significant events were captured by
Landsat 5, such as 1986s Chernobyl incident, 2004s
devastating tsunami in Southeast Asia, etc. [58] Landsat 5 set a
Guinness world record for Longest operating Earth
observation satellite. Landsat 5 Thematic mapper sensor data
is available on USGS websites for free. [59]
To assess short term (in a span of 10 and 13 years) and long
term (in a span of 23 years) shoreline change detection, cloud
free and low tide data of Landsat 5 TM were downloaded from
the USGS EarthExplorer website. Further characteristics of
satellite data (Landsat 5 TM) used in this study is summarized
in table 3. Boundaries of water bodies can be delineated more
accurately using fine resolution spatial data for example
IKONOS and SPOT, with spatial resolution of less than 1 m.
Although IKONOS and SPOT are commercial satellites and
their data is not available free of cost. However, all the Landsat
data in USGS archive is available online for free. The summary
of collected data is given in table 4.
Table 2 Length of Pakistan Coastline from different sources

Serial
#
Author/
Organization
Year
Length
of
Pakistan
Coastline
km
Length
of
Makran
Coastline
km
Length of
Sindh
Coastline
km
1
2
UNEP [34]
Quraishee [39]
Tariq [52]
Rashid [23]
1986
1986
2000
2012
825
990
___
___
___
___
Pakistan
Tourism
Development
Corporation
[53]
Jalal [54]
Wildlife of
Pakistan [32]
2006
1046
___
___
2008
2006
1050
1050
___
800
___
250
Pike [36]
Hafeez [55]
2014
2007
1050
700
350
Einfopedia
[56]
Encyclopedia
Britannica
[25]
ESCAP [21]
WWFPakistan [24]
Himalayan
Holidays [26]
2010
1100
771
329
2001
___
___
350
1996
2008
___
___
___
1000
370
___
2002
___
754
___
Jalal [33]
2008
1200
___
___
4
5
9
10
11
12
Proceedings
B. Data Processing
After data collection, data was processed to make it ready for
our analysis. As the study area was very large (i.e. about 990
km long) the study area was divided into two parts i.e.
Western Makran Coast (Path/Row: 155-42, 155- 43, 156-42
and 156-43) and Eastern Makran Coast (Path/Row: 154-42,
154-43, 153-42 and 153-43) (See figure 6). All the images of
Western and Eastern study area were geometrically corrected,
stacked and mosaicked using ERDAS Imagine 9.2. All the
collected and processed data, and result files of shoreline
change detection and erosion - accretion analysis were stored
in ArcGIS geodatabase. All images and shape files used in this
study were projected to Universal Transverse Mercator
(UTM) projection system, zone 41 and the spheroid and datum
were referenced to WGS1984.
Fig. 2 Some Glimpses of Study Area- Makran Coast, Pakistan
A. Data Collection
For this study, we utilized Landsat 5 TM images of the
study area (Path/Row: 156-42, 156, 43, 155-42, 155-43, 154-
C. Shoreline Identification
Water behaves like a semi-transparent medium for the
electromagnetic radiation, hence when EM radiation strikes
water surface, it gets reflected, transmitted or absorbed in
water. The spectral responses depend on the wavelength of the
radiation as well as the physical and chemical properties of the
water. When water is in the liquid form, it shows greater
reflectance in the visible region between 0.4m and 0.6m
however wavelengths beyond 0.7m are completely absorbed.
Water absorbs most of the energy in NIR and MIR
wavelengths and appears dark in these bands, in contrast to this
land and vegetation show higher reflectance in these
299
Proceedings
Fig. 5. Spectral Profiles of Land, Water and Vegetation drawn in Erdas

Imagine 9.2 from Landsat 5 TM image
be combined to make color composites. This will highlight

certain features in distinctive colors [65].
The initial and crucial step in the shoreline change
detection is the selection of shoreline proxy and then its
identification. Water boundary delineation is one of the
applications of Remote Sensing. In this study, we considered
High Water Line (WHL) as a shoreline proxy. The WHL is
defined as the intersection of land with water surface at an
elevation of high water [66]. High water line has been used in
several studies for shoreline delineation [4, 10, 13, 41, 43, 44].
To identify the high water line (WHL), the mid infrared (band
5) and green (band 2) band ratio index methodology was
adopted.
Fig. 3 Materials and Methods
Fig. 4. Data Processing
wavelengths and appear bright in these bands. This property of

Water is used to differentiate it from land and vegetation in
multi-spectral imagery [7, 9, 62-64]. Figure 5 shows the
spectral profiles of land, water and vegetation drawn in
ERDAS Imagine 9.2 from a Landsat 5 TM image.
Figure 7 shows images of Miani Hor -a part of Makran
Coast- in different bands of the Landsat 5 TM. In first three
bands the contrast between water and other surface features are
not very remarkable. On the contrary, the IR bands (bands 4
and 5) show a sharp contrast between them as water reflects
very less in the IR region of the EMR spectrum. Among all the
bands, band 5 gives the most outstanding contrast between
water and other surface features. However, different band
combinations have their own tendency to differentiate land and
water. Figure 7 shows that Mid IR and Green band ratio is very
suitable for shoreline delineation.
Band ratio is a digital image processing technique in which
DN value of one band is divided by that of any other band. It is
used to enhance the contrast between different features. It also
eliminates the shadowing effects [63]. Band ratio has its
application in several fields such as minerals, soil, vegetation,
water bodies etc. [64]. Three different band ratio images can
D. Shoreline Extraction
Band ratio was calculated for every image, then the high
water line was extracted by manual digitization on 1:50,000
scales in ArcGIS 10.3. This scale was kept consistent to in all
digitization work throughout the study to maintain the
consistency. The temporal movement (1987-2010) of shoreline
is shown in figure 8 and 9.
E. Shoreline Length Measurement
The length of shorelines (extracted from band ratio images
of 1987, 2000 and 2010) were measured using calculate
geometry function in ArcGIS 10.3.
F. Shoreline Change/Movement
Shoreline change was analyzed by overlay technique.
Extracted shorelines of 1987, 2000 and 2010 of eastern and
western Makran coast were overlaid on each other to visualize
the shoreline movement.
G. Erosion and Accretion Measurement
Erosion is the transportation process of soil or decomposed
rock material by natural forces (such as storms, flooding,
waves and tides) while accretion is a slight and unnoticeable
accumulation of soil or decomposed rock by natural
phenomenon. Coastal accretion occurs when the alluvion
deposit upon the shore. [68]
300
Proceedings
Fig. 6. Study Area- Makran Coast, Pakistan
Erosion and accretion analysis for short time spans (i.e.

1987-2000 and 2000-2010) and long time span (i.e. 19872010) for Western and Eastern Makran Coast were done using
ArcGIS 10.3. The Features to polygon tool was applied to 1987
and 2000 shorelines to identify the stable and unstable areas.
This tool gave unstable areas as polygons. From these unstable
areas (polygons) the advanced (accretion) and retreat (erosion)
areas were identified and their areas were calculated using
ArcGIS 10.3. The same steps were followed to do the accretion
and erosion measurements for another short time span (i.e.
2000-2010) and a 23 year long time span (i.e. 1987-2010). To
better understand the results were normalized (See table 6-8).
Figures 10-17 show the developed accretion and erosion maps.
H. Visualization
GIS has provided us a new way of looking and
communicating our results. Visualizing results for better
understanding has become very common these days. All the
results of this study were visualized using ArcGIS 10.3.
The division of study area in eastern and western part,
movement of shoreline and accretion - erosion in different time
spans are shown in figures 6, 8 and 9, and 10-17respectivelyy.
Coastline paradox is a globally accepted phenomenon. The
length of coastline changes when measured from remote
sensing data of different resolution or when measured on
different scales. It also changes when measured on the same
scale from remote sensing data of the same resolution at
different time interval. The reason behind this temporal
variability is the change in shape of coastline associated with
erosion and accretion. Table 5 shows the length of Makran
coastline measured in different years from Landsat TM images.
However the length of coastline was measured on the same
scale from remote sensing data of same spatial resolution
images, the length of coastline was found different in different
years due to change in coastline shape with associated erosion
and accretion. Spatio-temporal shoreline movement and

accretion - erosion are shown in figures 8-17.
A. Accretion - erosion along Western Makran Coast
The western Makran coast accretion - erosion analysis
results are presented in table 6. The accretion - erosion analysis
result is geographically presented in figure 10.
1) Accretion and Erosion between 1987-2000
The accretion - erosion analysis shows during 1987 to 2000
(i.e. 13 year time span) about 9.4 sq. Km land was eroded
while 12.6 sq. Km land area was advanced. Hence the total
change was 3.2 sq. Km land advancement. During 1987-2000
the accretion rate per year was about 1 sq. Km, while the
erosion rate was 0.7 sq. Km which was less than accretion rate.
Hence, in this short time span (1987-2000) accretion was more
dominant than erosion along the western coast and the land
advanced at a rate of 0.3 sq. Km per year.
The accretion - erosion analysis shows during the short
time span of 10 years (i.e. 2000- 2010) about 11 sq. Km land
was eroded while land area was advanced by 4.8 sq. km along
western Makran coast. Hence the total change was 6.2 sq. Km
erosion. During 2000-2010 the accretion rate per year was
about 0.4 sq. Km, while the erosion rate was 1.1 sq. Km, which
was quite higher than accretion rate. Hence the erosion was far
dominant than accretion in this time span and the Western
Makran coast receded at the rate of 0.7 sq. km per year.
The accretion - erosion analysis shows during the long time
span of 23 years (i.e. 1987- 2010) about 13.6 sq. Km land was
eroded while land was advanced by 9.2 sq. km. Hence the total
change was 4.4 sq. Km erosion. During 1987-2010 the
accretion rate per year was about 0.4 sq. Km, while the erosion
rate was 0.6 sq. Km which was greater than accretion rate.
Hence the erosion was more dominant than accretion in this 23
year long time span (1987-2010) and the coastal area of
Western Makran was eroded at the rate of 0.2 sq. Km per year.
301
Proceedings
Fig. 7. Landsat 5 TM Images of Miani Hor-a part of Makran Coast- in different Spectral bands, band combinations and band ratio using ERDAS Imagine 9.2
4) Accretion - erosion Analysis Results on Local Scale

a) Gwadar
Gwadar (250735N 621921E) is a city located on the
southwestern coastline of Pakistan. Gwadar was nominated as
the winter capital of Balochistan in 2011. Gwadar port located
at the mouth of the Persian Gulf. It is a third deep sea port of
Pakistan, inaugurated to increase the port capacity of Pakistan
[69]. The erosion - accretion analysis shows that Gwadar coast
has faced a noticeable erosion problem. During 1987-2000
about 4.5 sq. Km land was eroded at Gwadar port and no
noticeable accretion took place while 1.5 sq. Km land was
eroded and 0.5 sq. Km land was advanced at Western Gwadar
bay. During 2000-2010 about 1.6 sq. Km land was advanced at
Gwadar port while no noticeable erosion took place
meanwhile, no noticeable erosion and accretion took place at
Eastern or Western Gwadar bay. In long time span of 23 years

(1987-2010) about 4 sq. Km land was eroded at Gwadar port
while 1.5 sq. Km land was advanced meanwhile 1.2 sq. Km
land at western and 0.4 sq. Km land at eastern Gwadar bay was
eroded while no noticeable accretion took place (See figure
11).
b) Pasni
Pasni (2516N 6328E) is a moderate size town and
fishing port located at the Makran coast [70]. During the time
span of 13 years (1987-2000) about 0.2 sq. Km land was
eroded while 4.5 sq. Km land was advanced at Pasni and its
vicinity. During the short time span of 10 years (2000-2010)
about 0.3 sq. Km land was advanced while 5.8 sq. Km land
was eroded at Pasni and its vicinity which is greater than the
accretion took place during 1987-2000. During the long time
302
Proceedings
Table 3 Landsat 5 TM and MSS Sensor Characteristics

Satellite
Sensors
Landsat
5
Swath
(Km)
Scene size
(Km)
Temporal
Resolution
Spectral
180
180 x 170
16 days
MSS
TM
185
170 x 183
16 days
Table 4 Summary of Collected Data

Satellite
(Sensor)
Landsat 5
(TM Sensor)
Path/Row
153-42,
154-42,
155-42,
156-42,
153-43
154-43
155-43
156-43
Year
Tide
Phase
1987
2000
2010
Low
Table 5 Makran Coastline Length measured from Landsat 5 TM images

Year
Scale
Spatial
Resolution
m
2010
2000
1987
1:50,000
30
Western
Makran
coast
length
Km
335
337
344
Eastern
Makran
coast
length
Km
634
661
667
Makran
coast
total
length
km
969
998
1011
Shoreline
proxy
WHL at
low tide
span of 23 years (1987-2010) about 0.5 sq. Km land was

eroded as well as 0.5 sq. Km land was advanced which showed
a balance in erosion and accretion but at a long time span (See
figure 12).
a) Jiwani
Jiwani (2503N 6144E) is a town and commercial port. It
is located near the Pak-Iranian border [71]. The accretion erosion analysis shows that the coastal area of Jiwani and its
vicinity is under influence of natural forces as well human
induced factors which have been causing coastal erosion.
During 1987-2000 about 5.2 sq. Km land was advanced while
no considerable erosion took place. During 2000-2010 about
4.8 sq. Km land was eroded (near about accretion took place
during 1987-2000) while only 0.8 land was advanced at Jiwani
and vicinity. In 23 year long time span (1987-2010) about 5.2
sq. Km land was advanced while 2.4 sq. Km land was eroded
(See figure 13).
Band
Band 4 Green
Spectral
Coverage
(m)
0.5 0.6
82
Band 5 Red
0.6 0.7
82
Band 6 Near IR
0.7 0.8
82
Band 7 Near IR
0.8 1.1
30
Band 1 Blue
0.45 0.52
30
Band 2 Green
0.52 0.6
30
Band 3 Red
0.63 0.69
30
Band 4 Near IR
0.76 0.90
Spatial
(Meters)
82
30
Band 5 Mid IR
1.55 1.75
120
Band 6 Thermal
10.40 12.50
30
Band 7 Mid IR
2.08 2.35
B. Accretion - erosion along Eastern Makran Coast

The western Makran coast accretion - erosion analysis
results are given in table 7. The accretion - erosion analysis
visualization is shown in figure 14.
while 12 sq. Km land area was advanced at the east Makran
coast. Hence the total change was 3.5 sq. Km accretion. During
1987-2000 the accretion rate per year was about 0.9 sq. Km,
while the erosion rate was 0.7 sq. Km which was not exactly
same, but close to the accretion rate. Hence, in this short time
span there was a balance in the accretion and erosion to some
extent. In this (1987-2000) thirteen year period the eastern
Makran coast was advanced at a rate of 0.3 sq. Km per year.
time span of 10 years (i.e. 2000- 2010) about 5.5 sq. Km land
was eroded while land was advanced by 16 sq. km at eastern
Makran coast. Hence the total change was 10.5 sq. Km
accretion. During 2000-2010 the erosion rate per year was
about 0.5 sq. Km, while the accretion rate was 1.6 sq. Km,
which was quite higher than erosion rate. Hence the accretion
was far dominant than erosion in this time span and land
advanced at the rate of 1.1 sq. km per year.
eroded while land was advanced by 17.5 sq. km at eastern
Makran coast. Hence the total change was 10.7 sq. km
erosion. During 1987-2010 the accretion rate per year was
about 0.8 sq. Km, while the erosion rate was 0.3 sq. Km which
was lower than accretion rate. Hence the accretion was quite
dominant than erosion in this 23 year long time span (19872010) and the coastal area of Eastern Makran was advanced at
the rate of 0.5 sq. Km per year.
303
Proceedings
Fig. 8. Western Makran Coast Temporal Shoreline Movement 1987-2010
Fig. 9. Eastern Makran Coast Temporal Shoreine Movement 1987-2010
4) Accretion - erosion Analysis Result on Local Scale

a) Hingol River
Hingol River (25 23N 65 28E) is located in the Gwadar
district of Balochistan. It is the longest river in Balochistan
(about 560 km long). Unlike the majority of streams in
Balochistan, which flow only in rare rain, Hingol River flows
all year long [72]. The erosion - accretion analysis of mouth of
the Hingol River shows that during 1987-2000 about 1.3 sq.
Km land was eroded there while 1.2 Sq. km land was advanced
in its vicinity. During 2000-2010 about 0.4 sq. Km land was
advanced while only 0.2 Sq. km land was eroded at the mouth
of Hingol River and about 0.6 Sq. km in the vicinity. In long

time span of 23 years (1987-2010) about 0.2 sq. Km land was
eroded while 0.3 sq. Km land was advanced at the mouth of the
Hingol River while in its vicinity about 0.6 Sq. km land
advancement took place and 1.2 sq. km land was eroded. (See
figure 15).
b) Ormara
Ormara (251253N 64382E) is an old coastal town
located in the Gwadar district of Balochistan. Ormara has a fish
harbor and a port [73]. The erosion - accretion analysis of
Ormara shows that during 1987-2000 about 0.6 sq. Km land
304
Proceedings
Fig. 10. Western Makran Coast Accretion Erosion 1987-2010
was eroded while 4 Sq. km land was advanced at Ormara.

During 2000-2010 about 1 sq. Km land was advanced at
Ormara port while only 1.5 Sq. km land was eroded. In long
time span of 23 years (1987-2010) about 0.5 sq. Km land was
eroded at Ormara port while 3.7 sq. Km land was advanced.
(See figure 16).
c) Islands in Study Area

A piece of sub-continental land that is surrounded by water
is known as the Island. Islands are classified into two main
types i.e. continental islands and oceanic islands. They may be
artificial as well [74]. Pakistan has several natural islands off
the Sindh and Balochistan (Makran) coasts. The islands off the
Makran coast are Malan island (a volcanic mud island that
305
Proceedings
Fig. 11. Gwadar Accretion Erosion 1987-2010

appears and disappears), Zalzala island (appeared offshore
Gwadar followed by an earthquake of 7.7 magnitude in
September 2013) and the Astola island [75].
ASTOLA ISLAND
Astola Island (25721N 635051E) is Pakistans largest
offshore island. It is also known as Haft Talar Island (Island of
seven Hills) with the highest point 246 feet above the sea level
[76]. It is about 6.7 km long and with a maximum width of 2.3
km and 6.7 Sq. km approximated area [76]. However the area
of Astola Island estimated in this study was 2.2 sq. km with
shoreline length 10.2 km.
This study reveals that during short (i.e. 1987-2000 and 20002010) and long time (1987-2010) spans, very nominal
accretion and erosion took place along the shoreline of Astola
Island. During the short time span of 13 years (i.e. 1987-2000)
only 0.13 sq. km land was advanced while 0.02 sq. km land
was eroded. During ten year time span (2000-2010) only 0.04
sq. km land was advanced while 0.01 sq. km land was eroded.
During the long time span of 23 years (i.e. 1987-2010) 0.1 sq.
km land advanced while only 0.04 sq. km land was eroded.
d) Lagoons in Study Area
The lagoon is a shallow water body separated from the
larger water body by reefs and small islands [77]. Lagoons are
classified into coastal and oceanic lagoons. Coastal lagoons are
further classified into three main types i.e. choked, leaky and
restricted [78]. Lagoons are found in a great range of sizes, the
size of a smallest lagoon is about a hectare while the largest
(i.e. Lagoa Patos in Brazil) is about 10,000 km2. Usually the
length of the lagoon is greater than its width [79].
Lagoons are typical coastal features found parallel to the
coastlines all around the world say Glenrock lagoon in
Australia, Lagoa dos Patos lagoon in Brazil, Apoyo lagoon
natural reserve in Nicaragua, Blue lagoon in Turkey, Venetian
lagoon in Veneto, Miani Hor and Khor Kalmat along Makran
coast, Pakistan [77, 80, 81].
306
Proceedings
Fig. 12. Pasni and Vicinity Accretion Erosion 1987-2010
KHOR KALMAT
Khor kalmat lagoon (25 24N 64 06E) is located along
the eastern Makran coast. Basol River flows southward in
the Gwadar District. It drains a deserted Makran region, with
its river mouth at Khor Kalmat lagoon [81].
MIANI HOR
Miani Hor (25 32N 66 12E) is a swampy lagoon and a
Ramsar wetland, located in Lasbela district [80].
Both Miani Hor and Khor Kalmat were not included in the
accretion - erosion analysis of this study as lagoons are
extremely dynamic ecosystem [78]. Further, they have their
own dynamics different from that of other coastal features [79].
A study in which 7 year beach profile data was analyzed,
showed that the lagoon shorelines were far dynamic than the
ocean shorelines [82]. The flow of water in lagoons is slow and
sluggish and the flushing time depends on the type of lagoon
[78].
C. Accretion - erosion along Makran Coast

The accretion - erosion analysis results of western and
eastern Makran coast were combined to understand the
accretion - erosion scenario of entire Makran coast.
D. Accretion and Erosion between 1987-2000
while land was advanced by 24.6 sq. Km at Makran coast.
Hence the total change was 6. 7 km accretion. During 19872000 the accretion rate per year was about 1.9 sq. Km, while
the erosion rate was 1.6 sq. Km which was not exactly same,
but close to the accretion rate. Hence, in this short time span
there were balance in the accretion and erosion to some extent.
However, in this (1987-2000) thirteen year period the Makran
coastal area advanced at a rate of 0.5 sq. Km per year.
307

time span of 10 years (i.e. 2000- 2010) about 16.5 sq. Km land
was eroded while land was advanced by 20.8 sq. km at Makran
coast. Hence the total change was 4.3 sq. Km accretion. During
2000-2010 the erosion rate per year was about 1.7 sq. Km,
while the accretion rate was 2 sq. Km, which was higher than
erosion rate. Hence the accretion was more dominant than
erosion in this time span and land advanced at the rate of 0.4
sq. km per year.
eroded while land was advanced by 26.7 sq. km at Makran
coast. Hence the total change was 6.3 Km erosion. During
1987-2010 the accretion rate per year was about 1.2 sq. Km,
Proceedings
while the erosion rate was 0.9 sq. Km which was lower than
accretion rate. Hence the accretion was quite dominant than
erosion in this 23 year long time span (1987-2010) and the
coastal area of Makran was advanced at the rate of 0.3 sq. Km
per year.
IV. CONCLUSIONS
It is evident from this study that RS and GIS are very
efficient and cost effective in the coastline length
measurement, island area estimation, shoreline change
detection and erosion - accretion analysis. Results show that
the Makran shoreline has changed noticeably from 1987 to
2010 due to dynamical processes such as tides, waves, strong
wind, river discharge, river sediments, SLR and human
induced factors as well.
Fig. 13. Jiwani and Vicinity Accretion Erosion 1987-2010
308
Proceedings
Fig. 14. Eastern Makran Coast Accretion Erosion Analysis 1987-2010
During 1987-2010 erosion was dominant along the western

Makran as it was eroded at the rate of 0.2 sq. km per year,
while accretion was dominant along the Eastern Makran coast
as it was advanced at the rate of 0.5 sq. km per year.
During the long time span of 23 years (1987-2010) Gwadar
port was eroded 4 sq. km which was quite higher than that of
Jiwani i.e. 2.4 sq. km. However Ormara, Pasni and Astola
Island were relatively stable places.
As the erosion process is far dominant than accretion along
western Makran coast and some areas of the eastern Makran
coast, it poses a major threat to coastal communities, properties
and habitats. A shoreline change detection should be done on a
regular basis and integrated in sustainable coastal resource

management and environment conservation policies.
The focus of this study was more towards the use of GIS
and remote sensing in shoreline change detection and accretion
- erosion analysis and coastline length measurements than the
natural process and human induced factors which brought these
changes. It is highly recommended to extend this study, to
relate the local conditions which are responsible for these
changes and come up with solutions to reduce the effects of
human induce factors while deal with the natural processes for
the sustainable coastal resource management.
309
Proceedings
Table 6 Western Makran Coast Accretion - Erosion Analysis

Time
Period
Time
Span
years
Accretion
Sq. Km
Erosion
Sq. Km
Change
in area
Sq. Km
Normalized Results
Comments
Accretion /year
Sq. Km/ year
Erosion /year
Sq. Km/ year
Change /year
Sq. Km/ year
19872000
13
12.6
9.4
+ 3.2
0.7
+0.3
20002010
10
4.8
11
-6.2
0.4
1.1
-0.7
19872010
23
9.2
13.6
-4.4
0.4
0.6
-0.2
Figure 15 Accretion Erosion analysis of mouth of Hingol River and vicinity (1987-2010)
310
Accretion
was little bit
higher
Erosion was
very
dominant
Erosion was
dominant
Proceedings
Table 7 Eastern Makran Coast Accretion - Erosion Analysis

Time
Period
Time
Span
years
Accretion
Sq. Km
Erosion
Sq. Km
Change
in area
Sq. Km
Normalized Results
Comments
Accretion/year
Sq. Km/ year
Erosion/year
Sq. Km/ year
Change/year
Sq. Km/ year
19872000
13
12
8.5
+3.5
0.9
0.7
+0.3
20002010
10
16
5.5
+10.5
1.6
0.5
+1.1
19872010
23
17.5
6.8
+10.7
0.8
0.3
+0.5
Figure 16 Accretion Erosion Analysis of Ormara port (1987-2010)
311
Accretion was
greater than
erosion
Accretion was
quite dominant
than erosion
Accretion was
dominant
Proceedings
`Table 8 Makran Coast Accretion - Erosion Analysis 1987-2010

Time
Period
Time
Span
years
Accretion
Sq. Km
Erosion
Sq. Km
Change
in area
Sq. Km
Normalized Results
Comments
Accretion/year
Sq. Km/ year
Erosion/year
Sq. Km/ year
Change/year
Sq. Km/ year
19872000
13
24.6
17.9
+6.7
1.9
1.4
+0.5
20002010
10
20.8
16.5
+4.3
1.7
+0.4
19872010
23
26.7
20.4
+6.3
1.2
0.9
+0.3
Figure 17 Accretion Erosion Analysis of Ormara port (1987-2010)
312
Accretion
was little bit
higher than
erosion
REFERENCES
[1] O. S. Eludoyin, Oduore, T., Obafemi, A. A., "Spatio-Temporal
Analysis of Shoreline Changes in Bonny Island, Nigeria,"
Ethiopian Journal of Environmental Studies and Management,
vol. 5, 2012
[2] L. T. Elizabeth H. Boak, "Shoreline Definition and Detection: A
Review," Journal of Coastal Research, vol. 214, pp. 688-703,
2005.
[3] M. F. Jimmy O. Adegoke, Godstime James, Ganiyu Agbaje,
Temi E. Ologunorisa, "An Assessment of Recent Changes in the
Niger Delta Coastline Using Satellite Imagery," Journal of
Sustainable Development, vol. 3, 2010.
[4] F. J. G. Laura Del Rio, Javier Benavente, "Shoreline change
patterns in sandy coasts. A case study in SW Spain,"
Geomorphology, vol. 196, p. 15, 2013.
[5] S. Maiti, Bhattacharya, Amit K., "Shoreline change analysis and
its application to prediction: A remote sensing and statistics
based approach," Marine Geology, vol. 257, p. 13, 2009.
[6] G. A. scale Kuleli Tuncay, Karsli Fevzi, Dihkan Mustafa,
"Automatic detection of shoreline change on coastal Ramsar
wetlands of Turkey," Ocean Engineering, vol. 38, p. 9, 2011.
[7] S. Nayak, "Use of satellite data in coastal mapping," Indian
Cartographer, vol. 22, pp. 147-157, 2002.
[8] F. Murray, "Shoreline changes interpreted from multi-temporal
aerial photographs and high resolution satellite images: Wotje
Atoll, Marshall Islands," Remote Sensing of Environment, vol.
135, p. 11, 2013.
[9] S. Maiti, Bhattacharya, Amit K., "Shoreline change analysis and
its application to prediction: A remote sensing and statistics
based approach," Marine Geology, vol. 257, p. 13, 2009.
[10] R. R. Manik Mahapatra, Rajawat AS, "Shoreline Change
Monitoring Along The South Guajrat Coast Using Remote
Sensing and GIS Techniques," 2013.
[11] R. J. T. T. A. H. Naji, "Detection of Shoreline Change in ALThirthar Lake using Remotely Sensed Imagery and Topography
Map," Ibn Al- Haitham J. For Pure & Appl. SCI, vol. 24, 2011.
[12] T. Yanli, "The Application of GIS and RS for Coastline
Detection and Risk Assessment to Enhanced Sea Level Rise,"
Master of Science.
[13] R. S. a. C. M. Vedast Makota, "Monitoring Shoreline Change
using Remote Sensing and GIS: A Case Study of Kunduchi
Area, Tanzania," Western Indian Ocean, vol. 3, pp. 1-10, 2004.
[14] J. K. Gary A. Zarillo, and Vickie Larson, "A GIS Based Tool for
Extracting Shoreline Positions from Aerial Imagery
(BeachTools) Revised," ENGINEER RESEARCH AND
DEVELOPMENT CENTER VICKSBURG MS COASTAL
AND HYDRAULICS LAB, vol. No. ERDC/CHL-CHETN-IV73, 2008.
[15] L. J. Moore, "Shoreline mapping techniques," Journal of Coastal
Research, pp. 111-124, 2000.
[16] D. w. G. S. Raju Aedla, D Venkat Reddy, "Automatic Shoreline
Detection and Change Detection Analysis of NetravatiGurpurRiver mouth Using Histogram Equalization and Adaptive
Thresholding Techniques," Aquatic Procedia, vol. 4, p. 8, 2015.
[17] I. A. Dar and M. A. Dar, "Prediction of shoreline recession
using geospatial technology: a case study of Chennai Coast,
Tamil Nadu, India," Journal of Coastal Research, pp. 12761286, 2009.
Proceedings
[18] T. D. T. O. D. T. Oyedotun, "Shoreline Geometry: DSAS as a

Tool for Historical Trend Analysis."
[19] T. T. B. Tran Thi Van, "Shoreline change detection to serve
sustainable management of coastal zone in Cuu long estuary " in
International Symposium on Geoinformatics for Spatial
Infrastructure Development in Earth and Allied Sciences, 2008.
[20] E. C. F. Bird, The coast of Victoria: the shaping of scenery:
Melbourne Univ Pr, 1993.
[21] ESCAP, "Coastal environmental management plan for
Pakistan," Thailand, Bangkok ST/ESCAP/1360, 1996.
[22] T. M. A. Khan, D. A. Razzaq, Q.-U.-Z. Chaudhry, D. A. Quadir,
A. Kabir, and M. A. Sarker, "Sea Level Variations and
Geomorphological Changes in the Coastal Belt of Pakistan,"
Marine Geodesy, vol. 25, pp. 159-174, 2010.
[23] A. I. Ali Rashid Tarbrez, S. M. Ali, M. Tabrez, M Danish, "The
Issue of coastal erosion and accretion along the Makran coast of
Pakistan," p. 10, 2012.
[24] WWF- Pakistan. (2008, June). Coastal Areas. Available:
http://foreverindus.org/ie_ecosystem_coastal.php
[25] Britanica The Editors of Encylopedia. (2001, June). Makran |
region,
Asia.
Available:
http://www.britannica.com/place/Makran-region-Asia
[26] Himalyan Holidays. (2002, June). geography - baluchistan / The
Makran
Coast.
Available:
http://www.himalayanholidays.pk/baluchistan_makrancoast.htm
l
[27] L. F. Richardson, "The problem of contiguity: an appendix of
statistics of deadly quarrels," General systems yearbook, vol. 6,
pp. 139-187, 1961.
[28] J. Akintola. (2014, June). GIS Data and the Coastline Paradox ~
GIS Lounge. Available: http://www.gislounge.com/gis-datacoastline-paradox/
[29] G. Darkes. (2008, June). The British Cartographic Society >
How
long
is
the
UK
coastline?
Available:
http://www.cartography.org.uk/default.asp?contentID=749
[30] A. P. Hristo Stanchev, Margarita Stancheva, Anatoly Apostolov,
"Determination of the Black Sea Area and coastline length using
GIS methods and landsat 7 satellite images " Geo-Eco-Marina,
2011.
[31] L. Y. C. Y. C. K. M. T. H. C F Chen, "The Measurement of
Coastlines Using SPOT-5 High Resolution Satellite Images,"
2015.
[32] Wild Life of Pakistan. (2001, June). Introduction to Pakistan:
Section
5:
Coastline.
Available:
http://www.wildlifeofpakistan.com/IntroductiontoPakistan/coast
lineofPakistan.htm
[33] Jalal. (2008, June). PakistanPaedia - Ports of Pakistan.
Available: http://pakistanpaedia.com/mega/ports.html
[34] UNEP, "Environmental Problems of the marine and coastal area
of Pakistan " 1986.
[35] K. Z. Stephen PLeatherman, Bruce C. Douglas "Sea Level Rise
Shown to Drive Coastal Erosion " Eos, Transactions American
Geophysical Union, vol. 81, pp. 55-57, 2000.
[36] J. Pike. (2014, June). Pakistan- Navy. Available:
http://www.globalsecurity.org/military/world/pakistan/navyintro.htm
[37] C. P. Rajendran, K. Rajendran, M. Shah-hosseini, A. N. Beni, C.
M. Nautiyal, and R. Andrews, "The hazard potential of the
313
[38]
[39]
[40]
[41]
[42]
[43]
[44]
[45]
[46]
[47]
[48]
[49]
[50]
[51]
[52]
western segment of the Makran subduction zone, northern

Arabian Sea," Natural Hazards, vol. 65, pp. 219-239, 2012.
G. Hoffmann, M. Rupprechter, N. A. Balushi, C. Grtzner, and
K. Reicherter, "The impact of the 1945 Makran tsunami along
the coastlines of the Arabian Sea (Northern Indian Ocean) a
review," Zeitschrift fr Geomorphologie, Supplementary Issues,
vol. 57, pp. 257-277, 2013.
Q. G. S, "Feasibility Studies for the Extraction of Energy from
Current and Haliohydrography along Pakistan Coast,"
NIO,1986.
A. Shetty, K. S. Jayappa, and D. Mitra, "Shoreline Change
Analysis of Mangalore Coast and Morphometric Analysis of
Netravathi-Gurupur and Mulky-pavanje Spits," Aquatic
Procedia, vol. 4, pp. 182-189, 2015.
U. Natesan, Parthasarathy, Anitha Vishnunath, R., Kumar, G.
Edwin Jeba, Ferrer, Vincent A., "Monitoring Longterm
Shoreline Changes along Tamil Nadu, India Using Geospatial
Techniques," in International Conference on Water Resources,
Coastal and Ocean Engineering, 2015, pp. 325-332.
B. J. A. A Vittal Hedge, "Shoreline Transformation Study of
Karnataka Coast: Geospatial Approach," presented at the
International Conference on Water Resources, Coastal and
Ocean Engineering, 2015.
R. G. Richa Choudharay, and V. Sanil Kumar, "Shoreline
change detection from Karwar to Gokarna - South West coast of
India using remotely Sensed data," International Journal of
Earth Sciences and Engineering, vol. 6, pp. 489-494, 2013.
A. Das, "Integrated Land-Use Study and Shoreline Changes
along Shankarpur - Mandarmoni Coast line, West Bengal Using
Remote Sensing and GIS Technique," Indian Cartographer, vol.
33, p. 6, 2013.
L. AZAZ, "Using GIS and Remote Sensing to map coastline
changes of Wedam_Alsahel area, Batinah, Oman between 1998
and 2008," presented at the The 33rd Asian Conference on
Remote Sensing, Thailand, 2012.
S. P. Aekkarak FAIBOON, Sompratana RITPHRING,
"Monitoring Shoreline Change Using Geo-Informatics: A Case
Study at Chalatat Beach in Songkhla Province, Thailand," 2012.
R. T. Kumaravel.S, Gurunanam.B, Suresh.M, "Quantitative
estimation of shoreline changes using remote sensing and GIS:
A case study in the parts of Cuddalore district, East coast of
Tamil Nadu, India," International Journal of Environmental
Sciences, vol. 2, 2012.
X. Li, Damen, Michiel C. J., "Coastline change detection with
satellite remote sensing for environmental management of the
Pearl River Estuary, China," Journal of Marine Systems, vol. 82,
pp. S54-S61, 2010.
A. a. J. Kumar, K. S., "Long and Short-Term Shoreline Changes
Along Mangalore Coast, India," International Journal of
Environmental Research, vol. 3, pp. 177-188, 2009.
L. M. Klein Micha, "Monitoring changes in shoreline position
adjacent to the Hadera power station, Israel," Applied
Geography, vol. 26, p. 17, 2006.
K. J. P. Paul Shane Frazier, "Water Body Detection and
Delination with TM Data," Photogrammetric Engineering &
Remote Sensing, vol. 66, p. 7, 2000.
T. M. A. Khan and M. Rabbani, "Sea Level Monitoring and
study of sea level variations along Pakistan coast: A component
of integrated coastal zone menegement," Karachi, Pakistan:
National Institute of Oceanography, 2000.
Proceedings
[53] Pakistan Tourism Development Corporation. (2006, June).

GeographyPakistan.
Available:
http://www.tourism.gov.pk/geography_pakistan.htm
[54] jalal. (2008, June). PakistanPaedia - Geography of Pakistan.
Available: http://pakistanpaedia.com/land/geo_1.html
[55] H. Hafeez, "The Potential of Tsunami generation along KarachiI
and the Makran coast of Pakistan," Pakistan Journal of
Meteorology, vol. 4, p. 16, 2007.
[56] einfopedia. (2010, June). Coastline of Pakistan: Longest
Coastline Of Pakistan Information of the world. Available:
http://www.einfopedia.com/longest-coastline-of-pakistan.php
[57] NASA. (2006, June). Landsat 5 Landsat Science. Available:
http://landsat.gsfc.nasa.gov/?p=3217
[58] USGS. (2013, June). USGS Release: USGS Completes
Decommissioning
of
Landsat
5
Available:
http://www.usgs.gov/newsroom/article.asp?ID=3626#.VX30iEa
ulj9
[59] NASA. (2006, June). Where to Get Data Landsat Science.
Available: http://landsat.gsfc.nasa.gov/?page_id=2367
[60] T. V. R. D. Nagesh Kumar, "Remote Sensing Applications in
Water Resources," Journal of the Indian Institute of Science,
2013.
[61] D. Mitra, A. Mishra, T. Phuong, and R. Sudarshana,
"Application Of Remote Sensing And Geographic Information
System For Coastal Zone Management In Sagar Island, Bay Of
Bengal," International Archives Of Photogrammetry Remote
Sensing And Spatial Information Sciences, vol. 34, pp. 435-439,
2003.
[62] G. L. Smith and G. A. Zarillo, "Calculating long-term shoreline
recession rates using aerial photographic and beach profiling
techniques," Journal of Coastal Research, pp. 111-120, 1990.
[63] Esri. (2015, June). band ratio - GIS Dictionary. Available:
http://support.esri.com/en/knowledgebase/GISDictionary/term/b
and%20ratio
[64] H. E.-S. M.R. Mohammad, K. Sedeik, M. El Raey,
"Applications of Band Ratios identified by HHRRfor
recognition of surface units using TM data in Siwa depression,
Western Deser, Egypt," presented at the 22nd Asian Conference
on Remote Sensing, Singapore, 2001.
[65] M. Bay. (2001, June). Image Processing and Interpretation.
Available: http://fas.org/irp/imint/docs/rst/Sect1/Sect1_15.html
[66] S. M. Glenn, W. Boicourt, B. Parker, and T. D. Dickey,
"Operational observation networks for ports, a large estuary and
an open shelf," Oceanography-Washington DC-Oceanography
Society-, vol. 13, pp. 12-23, 2000.
[67] USGS. (2014, June). Landsat 5 History. Available:
http://landsat.usgs.gov/about_landsat5.php
[68] A. L. Shalowitz, Shore and sea boundaries: with special
reference to the interpretation and use of Coast and Geodetic
Survey data. vol. 2: US Department of Commerce, 1964.
[69] Wikipedia.
(2007,
July).
Gwadar.
Available:
https://en.wikipedia.org/wiki/Gwadar
[70] Wikipedia.
(2007,
July).
Pasni
(City).
Available:
https://en.wikipedia.org/wiki/Pasni_(city)
[71] Wikipedia.
(2014,
July).
Jiwani.
Available:
https://en.wikipedia.org/wiki/Jiwani
[72] Wikipedia. (2008, August). Hingol River. Available:
https://en.wikipedia.org/wiki/Hingol_River
314
[73] Wikipedia.
(2015,
August).
Ormara.
Available:
https://en.wikipedia.org/wiki/Ormara
[74] Wikipedia.
(2015,
August).
Island.
Available:
https://en.wikipedia.org/wiki/Island
[75] Wikipedia. (2015, August). List of Islands of Pakistan.
Available:
https://en.wikipedia.org/wiki/List_of_islands_of_Pakistan
[76] Wikipedia. (2015, August). Astola Island. Available:
https://en.wikipedia.org/wiki/Astola_Island
[77] Wikipedia.
(2015,
August).
Lagoon.
Available:
https://en.wikipedia.org/wiki/Lagoon
Proceedings
[78] S. Miththapala, Lagoons and Estuaries vol. 4. IUCN Sri Lanka

Country Office, Colombo.
[79] E. C. F. Bird, "Coastal lagoon dynamics," in Geomorphology,
ed: Springer Berlin Heidelberg, 1968, pp. 139-144.
[80] Wikipedia. (2015, August). Miani Hor. Available:
https://en.wikipedia.org/wiki/Miani_Hor
[81] Wikipedia. (2015, August). Khor Kalmat. Available:
https://en.wikipedia.org/wiki/Khor_Kalmat
[82] J. Harper, "Evaluation of Beach Profile Data From Betio and
Bairiki,
Republic
of
Kiribati,"
1989.
315
Proceedings
Influence Analysis of Minerals on Drinking Water

Quality Around River Jhelum
1*
Sadaf.Javed, 1Asad. Ali, 1Saleem.Ullah,
Razia. Rani2
Department of Space Science

Islamabad 44000, Pakistan
*nomi4all786@gmail.com
Department of Geography
Government College University
Lahore Pakistan
AbstractThe present study assessed the drinking water

quality around River Jhelum during present time. Three
hundred drinking water sample results were collected from
District Water Testing Lab, Public Health Engineering
Department, Govt. of The Punjab, in Jhelum which was based on
random sampling. These sample results were comprised of
chemical analysis of three different variables that show different
results with time and space. Geo-statistical analysis and
prediction methods such as variogram modeling and Krigging
were employed to investigate the spatial variation in
morphological structure around river Jhelum. The results had
been compared with WHO, US-EPA and Pak-EPA criteria of
drinking water quality criteria by water testing lab.
Keywordsspatial structure; variogram; anisotropy; krigging
I. INTRODUCTION
Water covers 3/4 of our earths surface area and is one of
the vital elements of life. Water is basic need of all forms of
life on earth. The water ratio in oceans & seas is ninety-seven
percent and the rest of three percent is mainly obtained from
surface water. Surface water includes rivers, canals, fresh
water lakes, streams and groundwater like well water and
borehole water [1]. Water is important for the survival of life
for every living organism. Human beings consume about two
thousand milliliters of water daily.
Water is used for drinking, agriculture sector, industries,
and domestic work [2]. According to World Water
Development Report estimates of 2011, seventy percent water
is used for agriculture purpose, twenty percent in industrial
sector and ten percent for domestic work. Further, more than
hundred million people have lack access to clean drinking
water, as estimated more than eighty percent deaths of children
is due to digestive diseases such as diarrhea (approximately
two million per year) that is caused by drinking contaminated
water [3]. According to the World Health Organization &
UNICEF Joint Monitoring Program for Water Supply and
Sanitation, and at least one eighty million people in the world
are drinking polluted water [4]. Water pollution issue has been
raised as one of the most severe problem for many developing
countries. Most of the river water in the urban areas of the
developing world is at the end point of wastes discharged from
the chemical industries [5]. Surface water is the most
defenseless to pollution due to easy availability for disposal of
polluted waters. Rivers play an important role in unification of
316
the public and industrial waste water and run-off from cropping
land [6]. Pollution of surface water with poisonous chemicals
of rivers and lakes with additional nutrients is of important
environmental fears for the entire world. Agricultural sector,
industrial belts, and urban activities are major sources of
substances and nutrients to marine ecosystems. On the other
hand, atmospheric deposition could be an important cause to
certain constituents [7].
Population growth along with fast urbanization, changing
lifestyles, and economic development has led to increasing
pressure on water sources [3]. Anthropogenic influences such
as urban, industrial and agricultural activities have increased
consumption of water resources in result of toxicities water. On
the other side, natural processes like changes in precipitation
contributions, erosion, and degradation of crustal materials
pollute surface water and harm its use for drinking, industrial,
agricultural, and recreational purpose [8]. Seasonal variations
in rainfall, surface overflow, interflow, groundwater flow and
forced in and outflows have a strong effect on river water
discharge system and, resulted in the absorption of toxins in
river water. Further, Public and industrial wastewater discharge
are continuously polluting water sources [9].
Aquatic impurity over natural processes is unimportant in
Pakistan. Inland sewage and built-up wastes are the major
contributors of water contamination. As a result major tributary
of river Chenab i.e. NullahAik has been tainted. Muddled
expansion and inadequate sewerage services have faster the
discharge of domestic liquid wastes without any management
[10].It is computable that almost 1/3 of the worlds inhabitants
use ground water for drinking. Throughout the 1990s, present
arsenic was found to be extensive in groundwater within the
America, Argentina, Taiwan, China, Hungary, and Vietnam.
River plain study conducted in 2005 disclosed that
Muzafargarh district in Pakistan is enriched in arsenic
concentration [11]. Water and Hygiene is highly unnoticed in
Pakistan. A great number of people do not have access to clean
drinking water as well as satisfactory sanitation systems and
toilets in Pakistan. According to statistical analysis of 2005,
around more than thirty eight million people did not have safe
drinking water source and more than fifty million people
lacked access to better hygiene facilities in Pakistan. It is
estimated that by the year 2015, millions of people will not
have clean drinking water and more than forty million people
will be deprived of tolerable hygiene facilities in Pakistan [12].
Water supply coverage with the help of piping and hand pumps
is around sixty-six percent in Pakistan. It is estimated that,
thirty percent of all diseases and forty percent of all deaths are
due to poor water quality in Pakistan. Water borne disease
name Diarrhea is stated as the primary cause of death in babies
and offspring in the country. On the other hand, every fifth
citizen suffers from disease caused by the contaminated water
[13]. Water quality assessment can be defined as the valuation
of physical, chemical and biological nature of water in relation
to natural quality, human effects and planned uses. A number
of water quality constraints have been developed to sum up
water quality records in an easily describable and understood
design [4]. In this paper we attempt to carry a geo-statistical
analysis of drinking water quality with three variables (pH,
TDS, and Calcium) which describe spatial structure, range,
directional changes, and anisotropy with the help of
variography & model fitting. Krigging, Krigging with drift
(trend removal) and Inverse Distance Weightage method for
interpolation were applied to predict surface area of Jhelum
city. Lastly, predicted data was cross validated for all three
variables.
II. STUDY AREA
Jehlum River is the primary source of water supply in Azad
Kashmir and Pakistan. The source of this river is spring VeriNag situated in Kashmir at the altitude of two thousand meters
and flows towards Pakistan [14].It is the largest and most
western of the five rivers of Punjab, and passes through
Sirinagar, Muzafrabad, Kohala, Jhelum, Gujrat, Khushab,
Mandibahudin, Sargodha, Chinnot, Mianwali and Jhang
Districts. It is a tributary of the Chenab River with the total
length of about five hundred & ten miles from which three
hundred & seventy nine miles are in Pakistan. Its tributaries are
Neelum, Kunhar, Kanshi, Poonch and Kahan rivers.
Proceedings
many contaminants are added in this river and then this water
is polluted. [14]. District Jehlum is lying in the western bank of
the river Jehlum. The water of River Jehlum is used for
hydroelectricity, irrigation and drinking purposes [15].
Figure 1 shows sample cites within Jhelum city which are
Bilal Town, Dhok Abdullah, Chishtian Mohallah, Bagh
Mohallah, Islam Pura, Islamia School Area, Jada, Railway
Colony, Iqbal Town, NayaMohallah, Machine Mohallah No.1,
Machine Mohallah No.2, Machine MohallahNo.3, Dhok
Jumma, Talianwala, Bahria Colony, Mojahidabad, Pera Ghaib,
Professor Colony, Abbas Pura, and Shadab Colony.
III. MATERIALS AND METHODS
City area of Jhelum district was selected which is situated

along western side of River Jhelum. The area is situated at 320
93` 31`` North Latitude, & 730 72` 63`` East Longitude [14].
All geo-statistical analysis was computed using different
packages & functions of R programing which is open source
software for all types of operating systems and final results
were produced using Krigging Interpolation. Mendely Desktop
and Jab-Ref were used for research paper citations and
referencing respectively.
Non spatial analysis was done before computing spatial
analysis. Non spatial statistics were calculated to find Mean,
Median, Standard Deviation, Variance, first & third quartiles,
skewness, kurtosis, minimum & maximum value for all three
variables. Further, histograms were plotted to check data
distribution for pH, TDS & Calcium. Before proceeding
towards spatial analysis, it was necessary to convert
coordinates into Universal Transvers Mercator Projection and
assigned World Geodetic System 1984 as reference system.
Histograms showed that data was not distributed equally. So, to
remove trend in data, de-trend function was used as written in
equation (1) below:
Z = 0+1 x+2 y+3 x2+4 y2+5 xy
Equation (1) is second degree polynomial. After de-trend,

variograms were computed to represent the spatial variation in
data set. In other words, variogram explains structural
continuity of all variables within the earth crust [16].The
variogram is defined as:
( )
( )
( )
* ( )
)+
In equation (2), (h): variogram, xi + h, xi : variables x at

location i, & i + h: lag vector, m(h): number of point pairs. The
variogram starts from zero and reaches its constant value called
sill (C). From where sill starts its called partial sill and denoted
as C1 and nugget starts from zero and denoted as C0, however
Fig. 1. Map of study area
River Jhelum has catchment area of 21,359 miles with the

11.85 MAF Annual average flows. Due to this long distance
317
when we talk about variogram sill we combine both partial sill

and nugget: C0 + C1 = C.
On theoretical grounds, variograms must be reduced to the
minimum value which is zero but when it reduced to zero it
means nugget effect C (0). Due to this condition variogram is
not enough for spatial variation analysis & suitable model must
be fitted on it [16]. For this purpose, four basic models Linear,
Exponential, Spherical, & Gaussian
(equations 3-6
respectively) were tested and one for each was selected which
had lowest standard error on each data value and fulfill
conditions of strong close-range dependency, very smooth at
each point with all point pairs, nugget, sill and range values.
These models differ from each other on vary basic
characteristics such as separations; semi-variances and lowest
standard error. These models are defined as:
( )
( )
( )
( )
)}
( ) }
) }
Proceedings
applied to these neighboring locations. The weights are

constrained to sum to unity; hence equation (7) is an unbiased
estimator [17].On the other hand, there are many statistical
methods which provide very accurate results in certain
situation such as the most comparable one to Krigging is the
inverse distance weighting method. But in our case, we use
only Ordinary Krigging interpolation method with most
appropriate models, to predict the values in the City. Figure 2
shows the flow chart of our study.
Geo-Statistical Analysis Method
Non-spatial Statistics
Mean, Median,
Variance, Quintiles etc.
Spatial Statistics
Variogram Analysis &

Models with lowest SE
Krigging as Interpolation
Method with Lowest RMSE
(6)
Cross Validation
In equations (3-6) all models have same parameters as, a is

denoted as range, h is lag distance, C is sill and exp is
exponential function in Gaussian and Exponential Model.
Anisotropic variograms show different ranges and sill values in
every direction (NS, EW, NE, NW, SE, SW, etc.). In other
words anisotropic variogram shows different spatial
continuities in different directions [16]. For this task, after
variography and model fitting, next step was to check
anisotropy in data set. Variogram maps, directional variograms
and rose diagrams were computed to check geometric and
zonal anisotropies for pH, TDS & Calcium variables. Finally,
Krigging as interpolation method was applied and cross
validated for all variables to counter check the resulted values
[17]. Krigging is the Best Linear Unbiased Predictor" (BLUP)
that satisfies a certain optimality criterion (so it's best" with
respect to the criterion) and provide minimum variance [19].
But it is only optimal" with respect to the chosen model and
the chosen optimality criterion [9]. The Krigging is based on
the weighted moving average [17].
( )
( )
In equation (7), x0 is estimation location z* is estimation of

variable at spatial location, z(xi) are observed values in this
case pH, TDS & Calcium values for each prediction from
nearest neighboring spatial locations & i values are weights
318
Fig. 2. Flow diagram of geo-statistical analysis method
IV. RESULTS AND DISCUSSION

A. Non Spatial Statistical Analysis of pH, TDS & Calcium
Table I shows summary of basic statistic behavior of all
variables such as mean, median, standard deviation, variance,
minimum & maximum values etc. Sample size of all variables
is 300 which are appropriate for spatial variogram analysis and
prediction. This non-spatial data makes basis for spatial
analysis. Minimum to maximum range of pH is 7.030 to 8.438,
TDS have 250 to 1390 & Calcium have 40 to 249 respectively.
Further histograms of pH, TDS & Calcium values describes
data distribution pattern is not normal for all variables.
TABLE I.
Data Attributes
Min
Max
Mean
Median
Variance
Std. Diviation
1st Q
3rd Q
Skewness
Kurtosis
Samples
Proceedings
BASIC STATISTICS OF PH, TDS & CALCIUM

Summary Statistics OF PH, TDS & Calcium
pH
TDS
Calcium
7.030
8.430
7.639
7.620
0.098
0.314
7.400
7.880
0.1834
-0.5808
300
250
1390
722.2
630
85139.68
291.787
490
911.2
0.6836
-0.6105
300
40
249
101.9
100
1421.872
37.707
74.5
125
0.9961
1.5911
300
Minimum and maximum values in histogram indicate that

there is not a single peak in histogram. Some outliers are also
there in pH & Calcium datasets. Due to this abnormality in data
there is anisotropic effect in variogram modeling for all cases.
So, there is a need to de-trend data set first before spatial
analysis with Interpolation methods.
Fig. 3. Linear Variogram Model for pH Variable
B. Spatial Statistical Analysis of pH, TDS & Calcium with

Interpolation Methods.
1) Variogram Analysis:
Resulted omnidirectional variograms with appropriate
models for pH, TDS & Calcium show correlation between
minerals with soil structure. Different models are fitted on the
bases of lowest standard error for each variable. Table II
presents S.E calculation for best model selection. Linear model
is most appropriate with 0.32 S.E value for pH, Exponential
Model with 297.20 S.E for TDS and Linear model with 39.01
S.E. Figures (3-5) show best appropriate variogram models for
all three variables. Point pairs of these variograms show some
kind of cyclic trend in data set. Red line in all variogram
figures is estimated model while green line is model fitting
line.
TABLE II.
Model Name
Linear
Exponential
Spherical
Gaussian
Fig. 4. Gaussian Variogram Model for Calcium Variable
STANDARD ERROR CALCULATION FOR BEST MODEL
Lowest S.E with Best Model for Each Variable

S.E for pH
S.E for TDS
S.E for Calcium
0.3284
0.3289
0.3286
0.3285
309.1631
297.2729
303.7566
305.7606
39.0119
39.2731
39.0272
39.0594
Fig. 5. Linear Variogram Model for TDS Variable
319
Proceedings
Omnidirectional variograms are not enough for analysis

rather directional variograms are needed to check anisotropic
trends in data. For this purpose directional variograms are
computed from 00 to 1650 with 150 angle tolerance. Figures (68) show anisotropic trends in data for pH, TDS and Calcium
variables respectively. These directional variograms clearly
show different point pair values in different directions. So, we
can say that all variables have anisotropic trends which mean
different values of minerals in soil have different values with
directions within earth crust in the city area of Jhelum around
River Jhelum.
Another way to check anisotropic trend is to compute rose
diagram and anisotropic map. Figures (9-10) show rose
diagram and directional trends with anisotropic map of pH
respectively.
Fig. 8. Directional Variograms of Calcium Variable
Fig. 6. Directional Variograms of pH Variable.

Fig. 9. Rose diagram of pH Variable
Fig. 10. Anisotropic Map of pH Variable
Fig. 7. Directional Variograms of TDS Variable.
320
Rose diagram and anisotropic map of pH clearly show

variation of data with different angles. Same is the case with
other two variables.If data is tilted in one side than both plots
show only that particular direction either 0 degree or 90 degree
but these trends explain level of all minerals is different
directions within earth crust. For more understanding, tables
(III-V) present variogram parameters with different angles.
Data of these tables are derived from directional variograms to
compare parameters of same data set with anisotropic effect.
Maximum values of nugget, sill, and range of three variables
are different with different angular directions with lowest S.E
based model fitting methods. As for pH variables minimum
nugget value is 0.03 at 105 degrees and maximum is 0.06 at
many angles. Minimum value of Sill is 0.13 at 60 degree and
maximum is 0.25 at 30 degree. Minimum Rang value is 350 at
90,105 & 165 degrees and maximum is 480 at 120 degrees. All
these variations lead to anisotropic effect in data values. Same
is the case with TDS & Calcium variables, sill, nugget and rang
are different with different directions. Only directional
variograms are not enough for analysis before computing
Krigging interpolation method. More statistical analysis of
variograms is required such as interpolated directional ranges
and anisotropic corrected variograms for all variables. All these
directional ranges are estimated directional coverage of
predicted surface area along with appropriate model fitting and
after that a final corrected variogram is required before
interpolation process. Figure (11) show interpolated directional
ranges for pH variable and which represents the same
methodology for other two variables. Lastly, one last corrected
variogram for each variable is required. Figures (14-16) show
variograms that are fitted with linear and exponential models.
These all corrected variograms are differing from above
presented on the basis of parameters. Its clearly shown in plots
that nugget, sill and range are not same as compared with
previously plotted omnidirectional variograms of same models.
TABLE III.
Direction
Omnidirection
0o
15o
30o
45o
60o
75o
90o
105o
120o
135o
150o
165o
TABLE IV.
Direction
Omnidirection
0o
15o
30o
45o
60o
75o
90o
105o
120o
135o
150o
165o
LINEAR MODEL VARIOGRAM PARAMETERS FOR PH
Proceedings
TABLE V.
LINEAR MODEL VARIOGRAM PARAMETERS FOR CALCIUM
Direction
Omnidirection
0o
15o
30o
45o
60o
75o
90o
105o
120o
135o
150o
165o
Variogram Parameters
Nugget
Sill
Range
20000
20000
25000
15000
15000
14000
13000
10000
10000
15000
15000
15000
15000
121000
121000
90000
70000
50000
114000
10500
100000
124000
129000
130000
115000
121000
370
370
250
200
200
150
150
200
250
300
500
300
400
Fig. 11. Interpolated directional ranges of pH variable
Nugget
Sill
Range
0.06
0.06
0.06
0.05
0.05
0.05
0.04
0.04
0.03
0.05
0.05
0.05
0.06
0.14
0.14
0.19
0.25
0.20
0.13
0.14
0.14
0.15
0.18
0.18
0.16
0.15
400
400
400
400
400
400
400
350
350
480
400
370
350
EXPONENTIAL MODEL VARIOGRAM PARAMETERS FOR TDS
Nugget
Sill
Range
1200
1200
1300
1300
1400
1480
1490
1490
1500
1100
1200
1200
1200
1550
1550
1400
1800
1900
1600
1800
1600
1600
2200
1800
2400
2000
150
150
100
100
100
100
100
130
130
200
200
350
480
Fig. 12. Corrected variogram of pH variable
321
Proceedings
Fig. 16. Isotropic variogram of TDS variable
Fig. 13. Corrected variogram of TDS variable
Fig. 14. Corrected variogram of Calcium variable
Fig. 17. Isotropic variogram of Calcium variable
2) Interpolation Method:
Krigging interpolation method is applied to all variables

and the results show that Krigging method is the best for
interpolation, as mentioned in methodology portion Krigging is
termed as best estimator so resulted plots are smooth and clear.
Moreover, Krigging also calculate standard error for each
predicted value over the surface. Final results are presented by
applying Ordinary Krigging interpolation method which clearly
describes observed and predicted values over entire surface
area of Jhelum city. Figure (18) is interpolation method for
observed and predicted values of pH variable along with
mapping of standard error for each predicted value. We applied
same methodology to acquire results with TDS and Calcium
Fig. 15. Isotropic variogram of pH variable
322
Proceedings
and get same type of models with little change thats why there
is no need to present those models here.
TABLE VI.
Lastly, Krigging cross validation is required to counter

check whether our methodology is corrected or not. Cross
validation is done with all 300 sample points along with their
respective variogram models. With the help of this cross
validation we are able to calculate MPE & RMSE values.
Tables (VI-VIII) present Krigging cross validation summary
for all three variables. Data attributes in validation reports
observed and predicted values along with residual values.
Observed and predicted values are close to each other that
prove our prediction method of Krigging is the best estimator.
Data Attributes
V. CONCLUSION
In nutshell, 300 sampling results of PH, TDS & Calcium
were collected from water testing lab situated in Jhelum city
and geo-statistically analyzed with the help of non-spatial and
spatial methods. With the help of non-spatial methods we
observed overall behavior of data by plotting histograms and
calculating basic statistics. After that spatial geo-statistical
analysis was applied with the help of variogram analysis and
Krigging interpolation method and found best estimator for
interpolation with the help of different plots and models.
Krigging is found best due to its smooth surface and S.E
calculation. Finally, Krigging cross validation process
rechecked the correctness of our Krigging method.
KRIGING CROSS VALIDATION OF PH VARIABLE
Minimum
Maximum
Mean
Median
1st Q
3rd Q
MPE
MPE/Mean
RMSE
RMSE/STD
RMSE/IQR
TABLE VII.
Minimum
Maximum
Mean
Median
1st Q
3rd Q
MPE
MPE/Mean
RMSE
RMSE/STD
RMSE/IQR
Residual
7.080
8.083
7.642
7.654
7.532
7.789
-0.0033
-0.0004
0.2517
0.8015
0.5244
-0.7449
0.8278
-0.0033
-0.0038
-0.1535
0.1325
-
Cross Validation Results

Observed
Predicted
Residual
250
1390
722.2
630
490
911.2
-
-834
1446.9
713.4
653
532.4
860.8
8.8686
0.0122
178.1268
0.6104
0.4228
-615.931
1464.050
8.869
-1.472
-60.488
61.605
-
KRIGING CROSS VALIDATION OF CALCIUM VARIABLE
Data Attributes
Minimum
Maximum
Mean
Median
1st Q
3rd Q
MPE
MPE/Mean
RMSE
RMSE/STD
RMSE/IQR
Fig. 18. Krigging interpolation of pH variable
Predicted
7.030
8.430
7.639
7.620
7.400
7.880
-
KRIGING CROSS VALIDATION OF TDS VARIABLE
Data Attributes
TABLE VIII.

Observed

Observed
Predicted
Residual
40
249
101.9
100
74.5
125
-
13.54
189.69
103.21
100.75
91.04
110.06
-1.3088
-0.0128
35.8101
0.9496
0.7091
-85.862
93.918
-1.309
-3.437
-26.856
25.972
-
ACKNOWLEDGMENT
The authors are indebted to Dr. Waqas Qazi who reviewed
the paper as a whole. Authors highly appreciated that Jhelum
water testing lab provided the data sampling results for this
research.
323
REFERENCES
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
J. McMurry, and R.C. Fay, Hydrogen, oxygen and water, 4th Ed. In:
K.P. Hamann (ed.). McMurry Fay Chemistry. Pearson Education, New
Jersey. pp.575-599, 2004.
A. Qayyum, K. Sulehria, Y. S. Mustafa, B. Kanwal, and A. Nazish,
Assessment of drinking water quality in Islampura , District Lahore,
( Local Report ), vol. 25, no. 2, pp. 359361, 2013.
R. Cidu, F. Frau, and P. Tore, Drinking water quality: Comparing
inorganic components in bottled water and Italian tap water, J. Food
Compos. Anal., vol. 24, no. 2, pp. 184193, 2011.
E. M. M. Wanda, L. C. Gulula, and G. PHiri, Determination of
characteristics and drinking water quality index in Mzuzu City, Northern
Malawi, Phys. Chem. Earth, vol. 5052, pp. 9297, 2012.
O.PHiri, P. Mumba, B.H.Z. Moyo, & W. Kadewa, Assessment of the
impact of industrial effluentson water quality of receiving rivers in urban
areas of Malawi, International Journal of EnvironmentalScience &
Technology, 2(3), 237244, 2005.
K. P. Singh, A. Malik, D. Mohan, and S. Sinha, Multivariate statistical
techniques for the evaluation of spatial and temporal variations in water
quality of Gomti River (India) - A case study, Water Res., vol. 38, no.
18, pp. 39803992, 2004.
Y. Ouyang, P. Nkedi-Kizza, Q. T. Wu, D. Shinde, and C. H. Huang,
Assessment of seasonal variations in surface water quality, Water
Res., vol. 40, no. 20, pp. 38003810, 2006.
V. Simeonov, J. a. Stratis, C. Samara, G. Zachariadis, D. Voutsa, a.
Anthemidis, M. Sofoniou, and T. Kouimtzis, Assessment of the surface
water quality in Northern Greece, Water Res., vol. 37, no. 17, pp.
41194124, 2003.
S. Shrestha and F. Kazama, Assessment of surface water quality using
multivariate statistical techniques: A case study of the Fuji River basin,
Japan, Environ. Model. Softw., vol. 22, no. 4, pp. 464475, 2007.
A. Qadir, R. N. Malik, and S. Z. Husain, Spatio-temporal variations in
water quality of Nullah Aik-tributary of the River Chenab, Pakistan,
Environ. Monit. Assess., vol. 140, no. 13, pp. 4359, 2008.
R. T. Nickson, J. M. McArthur, B. Shrestha, T. O. Kyaw-Myint, and D.
Lowry, Arsenic and other drinking water quality issues, Muzaffargarh
District, Pakistan, Appl. Geochemistry, vol. 20, no. 1, pp. 5568, 2005.
S. Anwer, S. a M. I. N. a K a u s a r, and K. Asghar, Factors affecting
drinking water quality and human health at household level in Punjab ,
Pakistan, Pakistan J. life Socail Sci., vol. 9, no. 1, pp. 3337, 2011.
S. Haydar, M. Arshad, and J. a Aziz, Evaluation of drinking water
quality in urban areas of Pakistan: A case study of Southern Lahore,
Pakistan J. Eng. Appl. Sci., vol. 5, no. July, pp. 1623, 2009.
S. Sarwar, F. Ahmad, and J. Khan, Assessment of the quality of Jehlum
River water for irrigation and drinking, vol. 23, no. 4, 2007.
M.J.Khan,S.Sarwarand R.A.Khattak, Evaluation of River Jehlum
water of heavy metals (Zn,Cu,Fe,Mn,Ni,Cd,Pb,and Cr) and it's suitabilty
for irrigation and drinking purposes at Districts Muzaffarabad(A.K),
Jour.Chem.Soc.Pk Vol.26,No.4,2004.
A.Hakimeh, G.Muhammad, R.L.Gholam,R., The applications of
geostatistical methods to prepare the 3D Petrophysical Model of oil
reservior,Open Journal of Geology,vol.3, pp.7-18, 2013.
James,R.,Charles E Glass, Use of geostatistics for accurate mapping of
earthquak ground motion,Geophysical Journal,pp.31-40,1997.
K.Chang, Introduction to geographic information systems,
6th ed,McGrawHill,New York, pp. 314-335, 2012.
R.Kerry,B.R.Ingram,P.Goovaerts & M.A.Oliver, How many samples
are required to estimate a reliable REML Variogram?, unpublished.
-------------------------------------------
324
Proceedings
Proceedings
Selection of the optimal interpolation method for

groundwater quality
Rida Batool, Khalid Mahmood, Hafiza Qimrah, Iqra Basit and Samia Rubab
Department of Space Science
University of the Punjab
Lahore, Pakistan.
Corresponding Author: ridabatool1214@gmail.com
AbstractThe study has been intended to find best possible
solution to predict the mixing of contamination with water and
alteration of the quality of groundwater, being used for drinking in
the mega city of Lahore. River Ravi and Lahore Canal are the two
main sources of groundwater recharge other than rain in Lahore so
ideally the areas near them should have same water quality but
unfortunately due to pollutants from industries, un-engineered
municipal waste dumping and improper sewerage system are spoiling
the underground aquifer. So our main focus is to find the trends in
spatial distribution of Total Dissolve Solids (TDS) concentration by
using different interpolation techniques. For this purpose TDS
concentration and coordinate information collected over 348
tubewells of the study area. Initially data has been analyzed by
different parameters such that skewness, kurtosis, standard deviation
and Q-Q plot for applying suitable transformation for normalizing the
data. With log transformation standard deviation reduced from
180.57 to 0.4388. Different types of Kriging stand out in optimizing
the best method among IDW, Spline, Radial Basis Function, Local
and Global Polynomials. Finally Ordinary Kriging concluded as the
more precise interpolation method for developing groundwater
quality index of TDS. For the selected method Average Standard
Error is 217.1m and Root Mean Square Error is 132.8m. Further
surface analysis in the form of contours and visual spatial appearance
has also supported the optimized interpolation method
Index Terms Ground Water Quality Index, Spatial Interpolation
Techniques, Surface Analysis, Contour analysis, Semivariogram.
I. INTRODUCTION
Earth surface is covered with 71% of water, 97.41% of that is
saline and only the remaining 2.59% is fresh water from
which useable fresh water is less than 1% [1]. Due to the
prompt growth of the population and immense increase in
industrialization the need for the fresh water has been
intensified [2]. According to experts till 2025, 52 nations are
going to suffer from severe shortage of drinking water [3].
Excessive evaporation and droughts also decreases the
availability of surface water.
Arid and semi-arid regions of the world like Pakistan, where
moon soon did not accomplished the need of surface water
and are likely to be scarce in near future [4]. 66% water is
supply in Pakistan is through piped networks and hand pumps
[5]. So, serious attempts are quite essential in regarding to
Groundwater resource management to restrict the reduction of
water for saving the future of next generation. World health
organization allots the guideline for drinking water which
decided how good or bad the quality of water for any motive
[6]. Groundwater quality index is the simplest tool which not

only used to measure the changes in water quality but also
provides the general analysis of the effects of water quality on
human life [7]. In most of the cases it is impossible as well as
cost effective to collect sample data from each location.
A review of previous studies concludes that geospatial
technologies help out in measuring pollutant concentration by
generating continuous surface in action of certain
geostatistical functions [8]. Spatial Interpolation is one of the
artistry tools used to access the ground water quality issues.
The motto of spatial interpolation is to generate a continuous
surface which is delineation of realism by using First law of
Geography (Toblers law) form basic of interpolation which
states that things close together are more similar than further
away [9].
A. Spatial Interpolation
Spatial data taken to be more reliable for Environment
management such as in Planning, risk assessment and decision
making. Spatial analysis can be carried out by no. of
interpolation techniques, which differs to eachother by rules or
differentiate the interpolators to the non-interpolators. On the
basis of the procedure and prospect, spatial interpolation
techniques describe in deterministic or stochastic method [10].
1) Deterministic Methods:
Usually deterministic methods uses spatial features of sample
data points to generate the Smooth surface.
a) Inverse Distance Weighted
IDW assign weights on the basis of distance between the
unknown location and the sampling points [11]. Usually IDW
expresses as:
(1)
Here the z is the predicted value,

is the value of known
point. And is the integer whose value ranging from 1 to n, is
the distance between the interpolated and measured point k is
the power parameter [12].
b) Global and Local Interpolation
In Global interpolation whole dataset is mapped out by
applying a single algorithm to generate continuous surfaces. In
local interpolators, it is optional to utilize either the single
325
function or multiple functions to interpolate the surface of

required area within limited window [10].
c) Spline
Spline uses complex mathematical function which interpolates
every point like bending a rubber sheet to pass through every
point. It reduces the curvature of the surface [13]. It is also
called Radial Basis Function. Basically Spline is consisted on
five different types of functions i.e. Completely Regularized
Spline, Spline with Tension, Multiquadratic Spline, Inverse
Multiquadratic Spline, and Thin Plate Spline. Commonly two
methods are used regularized and tension.
2) Stochastic Methods:
Second type of spatial interpolation methods is Geostatistical
methods and the most common geostatistical method is
Kriging. It is more efficient due to the statistical analysis of
the unpredicted spatial distribution of sample dataset. Kriging
facilitates with the selection of different algorithmic models of
semivariogram such as spherical, exponential, Gaussian,
Bessel etc [14]
a) Ordinary Kriging
Ordinary kriging is the basic and most probably usual type of
kriging. It surmises that trend has been unknown and constant
[15].
b) Simple Kriging
It presumes to be known the trend with constant random
variable [16].
c) Universal Kriging
Universal kriging suppose the existence of significant spatial
trend in sample points [17].When there is prominent trend in
data we use universal kriging which utilize stochastic method
to efficiently define the variation [18].
TDS belongs to the physical groundwater quality parameter
which composed of some of hydrocarbons and inorganic salts
such as bicarbonates, calcium, chlorides, magnesium,
potassium, sodium and sulfates. The origin of the dissolution
of such salts and organic components is very much dependent
upon the geological and climate variations in a given region
[19]. The basic aim of this study is to evaluate the best optimal
spatial interpolation method to generate the ground water
quality index for TDS
year 2013. Coordinate information of tubewells has been

collected by the field survey with GPS of accuracy of 3m.
Fig. 1. Study Area
Primarily the data was spread out using Excel. Further it

processed in the software i.e. ARC GIS where it analyze,
stored and display. General characteristics of sample data
comprised of TDS concentration are given below in the table 1
Table 1: Initial Statistics of TDS Sample Data
Obs.
No
Sum
Mean
Std.
Min
Max
Range
348
140307.1
391.9193
180.6993
137.7
1099.2
961.5
Statistical data of TDS ranges from 137.7 to 1099.2. Data

without transformation is more skewed than with
transformation. Histogram Analysis and Q-Q Plot help to
determine that which interpolation method effective for the
spatial data analysis. Transformation on TDS data has been
applied to bring data closer to Normal distribution. With log
transformation standard deviation reduced from 180.57 to
0.4388, as shown in the below table 2:
Table 2: Histogram Analysis of TDS Sample Data
Obs.
No
1
2

A. Study Area
Lahore hydrogeologically is a part of inter fluvial Bari Doab
which is surrounded by River Ravi to North West and Sutlej
and Beas River to the south-east [20]. Water and Sanitation
Agency (WASA) is providing the facility of water supply,
sewerage and drainage to almost 90% of the population of the
study area. WASA has installed 450 tube wells that supplies
720 mg of water over 531,336 connections [21].
Proceedings
Transf.
Mean
Median
Std.
Skewness
Kurtosis
No
Log
391.94
5.874
367.25
5.906
180.57
0.4388
1.2005
0.14261
4.6616
2.4138
Q-Q Plot analysis illustrates the comparison of approaching

normality by applying Log Transformation in Figure 2.
B. Data Collection
The dataset is consisting of in situ measured values of TDS
concentration level in water from 348 tube wells throughout the
Fig. 2.
326
Q-Q Plot; Before and After Applying Transformation
III. SELECTION OF INTERPOLATION METHOD
neighborhood into sectors [17].
Many interpolation techniques are available in GIS and each

method pertain (refer) to as model either simple or complex
models. A list of interpolation methods exists but difficulty is
to best regenerate the real surface from data points. In the
present study two basic classes of interpolation are used one is
Deterministic interpolation which are IDW, GPI, LPI and RBF
and the other is Geostatistical interpolation such as Ordinary
Kriging, Simple Kriging and Universal Kriging.
A. Deterministic Interpolation Methods:
IDW estimate the local variability of study area by using
linear weighted combination of sample points directly related
to the Optimize Power value. In case of TDS the optimized
power is 1 by including 6 neighborhood values at least with
anisotropic factor 1. Optimize power decreases the RMS up to
140.1. In Global Polynomial Interpolation, the power factor
for respective interpolation method has been 4 gives least
possible RMS value i.e. 157 by fitting smooth curve. Local
polynomial interpolation uses many polynomials to adjust so
that a smooth curve bring on. So the least possible RMS value
is 137 i.e. is interpolated by LPI. Completely Regularized
Spline estimated more closely to reality with an RMS value
equals to the 136.8 which has been more precise with least
possible error in all interpolation methods of Deterministic
Interpolation Techniques. Error statistics of deterministic
methods is given below in the table 3
Table 3: Error Statistics of Deterministic Interpolation Methods
Method
Proceedings
IDW
GPI
LPI
RBF
140.1
-1.51
-525.7
140.1
-610.0
385.8
995.8
157
-0.175
-60.99
156.98
-690.4
309.7
1000
137
-1.268
-447.5
137.3
-599.7
370.1
969.7
136.8
0.7318
254.7
136.8
-609.2
-343.1
953.3
3) Semivariogram:
Semivariogram and variogram are diagrammatic display of
autocorrelation with respect to the distance that similar as a
function between data points [24]. Choice of suitable
semivariogram for spatial data is time consuming [25]. To
point out the spatial variation, the consequential geostatistical
parameters are [11]
Nugget
Sill
Range
a) Nugget:
Nugget value is zero in ideal case but in real case variability
exist at short distance due to sampling error so the nugget
effect appears [26].
b) Sill:
The value where semivariogram flatten out, known as sill
[27].
c) Range:
Range help out to adjust the size of search window used in
spatial interpolation method [28]. If the correlation exists in
data points over a long range, variogram will retain least
amount of nugget effect and long range and vice versa [29].
4) Lag:
Separation distance between sample points, called lag [8].
Statistics
RMS
Mean
Sum
Std.
Min
Max
Range
5) Isotropic/ anisotropy semivariogram:

Isotropy exists when the spatial autocorrelation is identical in
an Omni direction, and in such case semivariogram model
does not depend upon the direction but only on the magnitude
of the distance [22].
IV. ADOPTED METHODS OF KRIGING
1) Search Neighborhood and Shape:

It used to stipulate distance and direction of measured points
to be used for prediction. Anisotropic factor can also be
changed by varying semi major axis and semi minor axis
according to the requirement and shape of neighborhood [17].
Adjustment of the range and minor range corresponding to the

log transformation helps to achieve uniform distribution.
Mathematical models as Circular Model, Spherical Model,
Exponential Model and Gaussian Model fitted on the
semivariogram and variogram data to acquire weights utilize
in kriging. Directional influence has applied by changing the
directional angle upto 45o to 52o to optimize the results of
TDS. In Simple Kriging, mean value estimated 5.874.
Transformation is also applied to interpolate with maximum
accuracy. In Universal Kriging an overriding trend in the data
and mean is known to the interpolator. Anisotropy applied by
changing the directional angle. Nugget value adjusted with
respect to the Partial sill for each mentioned Model. Following
are the Parameters needed to change according to their extent
to attain minimum RMS value in the given table: 4.
2) Sector Type:
It influence measured point in all direction by distributing each
A.
Statistical Accuracy Check Meters:
The exactness of interpolation depends on data acquisition and
B.
Geostatistical Interpolation
Kriging surmise that direction or distance between sampled
points reflects the spatial correlation [22]. Many accuracy
check methods are available, to compare the interpolation
method for the best fit surface [23]. When data is normally
distributed and stationary, Geostatistical interpolation give
effectives results. For optimal outputs certain parameters are
necessary to evaluate such as
327
Proceedings
Table 4: Parameters of Probabilistic Interpolation Methods

Method
Partial
Sill
0.182
Nugget
OK-C
Lag
Size
0.018
0.142
Minor
Range(m)
1.30
Major
Range(m)
0.214
OK-S
0.019
OK-E
0.019
0.084
0.132
0.06
0.216
0.096
0.115
1.00
OK-G
0.032
0.019
0.048
0.191
1.00
0.393
SK-C
0.008
0.131
0.062
0.04
0.039
SK-S
0.008
0.188
0.060
0.05
0.040
SK-E
0.007
0.124
0.075
0.03
0.076
SK-G
0.010
0.047
0.144
0.80
0.095
UK-C
0.029
0.090
0.193
0.40
0.212
UK-S
0.038
0.042
0.132
0.06
0.287
UK-E
0.018
0.095
0.040
0.30
0.070
UK-G
0.040
0.079
0.141
0.30
0.291
c) Mean Absolute Error:

Calculation of Mean absolute error sometimes preferred over
the RMSE as an evaluator because it is less reactive towards
extreme values. The formula of mean absolute error [37].
(3)
d) Average Standard Error:

We encounter overestimation of variability in prediction,
when average standard error is larger than Root mean square
prediction error and vice versa [24]. The formula [8]
(4)
Where
model applied in interpolation, the important issue is to select

an optimal method to fit on the sample points [30].
is kriged variance at location Xi
e) Mean Squared Error:

It is average difference between measuredvalue and predicted
value.
1) Validation:
Validation is much precise method, in which data divided into
two groups and remove one group, then interpolate through
that remaining data group and at the end test the removed
group values with the interpolated values [31].
(5)
2) Cross Validation:
Cross validation is quick, economical method for equating
prediction and measured values [32]. It eject temporary each
data point one by one and predict it with measured points.
Then take the difference of actual and predicted value to
determine the prediction accuracy [33].
Table 5: Comparison of Error Statistics of Different Interpolation

Methods
3) Error Plots:
The bias in between the predicted and measured value assist as
to evaluate the following statistics: mean error, root mean
square error, mean standardize error, root mean square
standardized error [34].
a) Root Mean Square Error:
Root mean square error is the under root of average square
difference between measured and estimated value. The
formula of RMSE [35].
(2)
Where SSE is the sum of error i.e. subtraction of predicted

over observed value and n is the number of pairs. The
accuracy of interpolation method can be determined by RMSE
and standardized RMSE [11].
b) Standardized Root Mean Square Error:
The standardized RMSE is only attainable for kriging but
RMSE is valid for every local method [36].
328
Method
RMSE
Mean
Std.
Range
ASE
MSE
IDW
140.1
-1.5
140.1
GP
157.0
-0.2
LP
137.0
RBF
RMSS
995.8
157.0
1000
-1.3
137.3
969.7
136.8
0.8
136.8
953.33
OK-C
132.8
10.5
132.4
1054.7
217.1
0.05
0.60
OK-S
137.2
07.3
137.0
992.7
177.3
0.06
0.76
OK-E
147.8
11.7
147.3
1119.9
185.6
0.05
0.76
OK-G
141.3
13.2
140.7
921.0
207.8
0.08
0.66
SK-C
135.6
00.1
135.6
996.2
133.4
0.01
0.97
SK-S
138.2
02.7
138.2
1024.1
139.7
0.02
0.95
SK-E
136.0
03.8
136.0
1003.8
150.2
0.04
0.87
SK-G
153.4
-0.3
153.4
989.3
161.5
0.03
0.95
UK-C
140.2
16.3
139.2
925.7
220.2
0.09
0.62
UK-S
136.7
07.1
136.5
1110.4
183.2
0.03
0.75
UK-E
135.0
-07.7
134.7
1019.3
103.0
1.31
UK-G
139.8
05.8
139.6
930.1
178.3
0.07
0.05
0.76
Proceedings
V. RESULTS AND DISCUSSIONS:
estimated by viewing prediction plots, on the basis of slope

nearer to 1. By checking all plots from the figure 4, it has been
concluded that ordinary kriging with circular model has more
slope than other plots approaches to 1.
Also ordinary and simple kriging with circular model produce
precise results. Global polynomial underestimating because it
has small y intercept value while local polynomial and
Gaussian simple kriging and universal kriging both are
overestimating as they have huge y intercept values.
The results concluded not only on the bases of interpolated

surfaces but also by using ultimate accuracy check elements.
Different types of Kriging stand out for the optimized methods
and none of the other deterministic techniques including IDW,
Global Polynomial, Local Polynomial and Radial Basis
Function etc. The results have specified that TDS analyzed to
be predicted well with a RMS error of 132.8 using Ordinary
Kriging with Circular Model. Other techniques and their Mean
value, Standard deviation, Range, Average standard error
(ASE), Mean standardized error (MSE) and Root mean
square standardized (RMSS) are mentioned in the following
table: 5.
C.
Semivariogram Analysis:
Circular ordinary kriging, Gaussian Simple kriging,
Exponential and Gaussian universal kriging semivariograms
depicting worse relationship, as showed in the figure 5 where
no correlation or least correlation exist between data points.
Circular ordinary kriging shows that as distance increasing
variance increasing gradually same as circular universal
kriging shows correlation to some extent but both have large
nugget. Simple kriging with Circular and Spherical curve have
been increasing asymptotically but most of the data lie on the
part of semivariogram which depict no any correlation.
A. Error Plots:
Ideal error should be zero but it obvious that interpolation
cannot be with 100% precision. In all displayed error plots in
Figure 3, methods show good or worse prediction values.
In GP, small range of error from 1.36 to -1.36 exists which
shows that it is efficient with its results. In GP, measured and
predicted values difference is equally increasing that are why
data points almost lie on line. In Gaussian simple kriging and
spherical universal kriging Error range is long so they possess
large errors. In IDW and spherical ordinary kriging data
spread immensely along line.
D. Surface Analysis:
Surfaces are the comfortable way to visualize interpolation,
which help us to predict the trend of parameters i.e. TDS in
fresh water. IDW and RBF both show nearly the same
prominent features such as Ravi, but IDW provide little spotty
features as from the given figure 6. Epicenter prominent well
in GP but does not indicate any variation. Similarly LP
displays that region with small spot without any variation.
Lahore Canal and River Ravi can be visualized up to some
extent in LP and GP. In ordinary Kriging spherical model
generates smoother surface but focal point described by
circular of ordinary kriging has been more prcised. Other two
models of Ordinary Kriging are edgy.
From each type of simple kriging, circular and spherical
illustrates a smooth surface whereas other two exhibits an
4) Contours:
The quality and reliability of contour maps helpful to find the
optimal method [38]. Texture of contours depend on many
factors such as no of neighbors influence the smoothness
(large no of neighbors) and roughness (least no of neighbors)
leading to the substantial variation in estimates [39].
B. Prediction Error Plots:

Prediction plots examined on the basis of y intercept and value
of slope. Dotted line is zero residual and blue line is regression
line in prediction plot. A precise interpolated surface can be
Fig. 3.
Error Plots Analysis of TDS
329
Proceedings
Fig. 4. Prediction Plots Analysis of TDS

edgy surface. Only exponential universal kriging stipulates
relatively smooth surface where midpoint illustrate better as
IDW and RBF both depict somehow a similar sort of contours

with small variations.
Fig. 5. Semivariogram Analysis of TDS

compare to the other three models of universal Kriging and
spherical of simple kriging. All demonstrates an area of
smooth area of over estimation in the lower right corner of the
image. All other display the variation of central and southern
spots.
E. Contour Analysis:
The analogy of the contours of probabilistic and deterministic
methods demonstrated in figure 7. Unwanted materials from
industries plus the change in the diverse interpolation methods
causes a variation in this ideal trend of River Ravi and Lahore
Canal. A comparison of different contour maps used to locate
some similar points so that trend of quality predicted well.
But RBF demonstrates a better variation in the epicenter.

There exists another loop of variation in the north eastern
aspect, but it appears better in IDW. GP generates an odd kind
of contours. Except of Exponential ordinary kriging every
other type refers the focal patch, Ravi and canal. In addition
all these three types provide another ring in the North-east
facet of the area. Each type of spherical kriging indicates the
epicenter except of Gaussian simple kriging along river Ravi
and Lahore canal with different variations. Although simple
spherical kriging elaborates a smooth of central contours. In
Universal kriging, Gaussian model illustrates an irregular
pattern of contours. Whereas exponential, circular and
exponential provide better visualization.
330
Fig. 6. Surface Analysis of TDS
Fig. 7. Contour Analysis of TDS
331
Proceedings
VI. CONCLUSION
This study concludes that the proper selection of the
interpolation method before computing water quality index for
TDS or any other water quality parameter is essential to have
the true inside of the local groundwater quality. The criteria of
selecting a suitable interpolation method rely on
characteristics of sample points, their mutual relationship,
attributes of interpolation methods and proficiency of
researcher and parameters interpolation sample size and
sample density etc. Different types of Kriging protrude for the
optimized methods as compared to any of the other techniques
including IDW, Global Polynomial, Local Polynomial and
Radial Basis Function etc. The results conclude that TDS is
analyzed to be predicted well with
minimum Average
Standard Error 217.1m and Root Mean Square Error 132.8m
using Ordinary Kriging with Circular Model. Accuracy of
different interpolation method can be check by distinct ways,
primarily RMSE, contour analysis, surface analysis, prediction
plots, error plots, semivariogram. The methods other than root
mean square error (RMSE) are visual and not quantitative
there for RMSE has been found as the better method for the
technique selection.
VII. REFERENCES
[1] Tahir, M. A., Rasheed, H., & Imran, S. (2010). WATER QUALITY
STATUS IN RURAL AREAS OF PAKISTAN. Islamabad: (PCRWR).
Retrieved from www.pcrwr.gov.pk.
[2] Ramakrishnaiah, C. R., Sadashivaiah, C., & Ranganna, G. (2009).
Assessment of Water Quality Index for the Groundwater in Tumkur
Taluk,Karnataka State, India. http://www.e-journals.net, 6 (2), 523-530.
[3] Kausar, S., Asghar, K., Anwar, S. M., Shaukat, F., & Kausar, R. (2011).
Factors Affecting Drinking Water Quality and Human Health at
Household. Pakistan Journal of Life and Social Sciences, 9 (1), 33-37.
[4] RizwanUllah, Malik, R. N., & Qadir, A. (2009). Assessment of
groundwater contamination in an industrial city, Sialkot, Pakistan.
African Journal of Environmental Science and Technology, 3 (12), 429446.
[5] Haydar, S., Arshad, M., & Aziz3, J. A. (2009). Evaluation of Drinking
Water Quality in Urban Areas of Pakistan: A Case Study of Southern
Lahore. Pak. J. Engg. & Appl. Sci, 5, 16-23.
[6] PCRWR (2007), Water Quality Monitoring Report Fifth monitoring
report (2005-6),I SBN 978-969-8469-18-4, Pakistan Council of
Research in Water Resources (PCRWR), www.pcrwr.gov.pk.
[7] Aenab, A. M., Singh, S. K., & Al-Rubaye, A. A. (2012). Evaluation of
Tigris River by Water Quality Index Analysis Using C++ Program.
Journal of Water Resource and Protection, 4, 523-527.
[8] Goroi, A. K., & Kumar, S. (2013). Spatial Distribution Analysis of
Groundwater Quality Index Using GIS: A Case Study of Ranchi
Municipal Corporation (RMC) Area. Geoinformatics & Geostatistics: An
Overview, 1 (2), 1-11.
[9] Tiengrod, P., & Wongseree, W. (2013). A Comparison of Spatial
Interpolation Methods for Surface Temperature in Thailand. Computer
Science and Engineering Conference (ICSEC), 2013 International, (pp.
174-178).
[10] Sterling, David L.(2003).A Comparison of Spatial Interpolation
Techniques For Determining Shoaling Rates of The Atlantic Ocean
Channel, Virginia Polytechnic Institute and State University. Unpublished
Masters thesis, Geography, Virginia Polytechnic Institute and State
University (pp.1-26).
[11] Su, P., Ting-xuan, L., Yong_dong, W., Hai-ying, Y., & Xi, L. (2009).
Spatial Interpolation and Sample Size Optimization for Soil Copper (Cu)
Investigation in Cropland Soil at County Scale Using Cokriging.
Agricultural Sciences in China, 8 (11), 1369-1377.
[12] Wang, F. (2006). Quantitative Methods and Applications in GIS (p. 43).
New York: Taylor and Francis group.
Proceedings
[13] Childs, C.,(2004). Interpolating Surfaces in ArcGIS Spatial Analyst, 3235. Retrieved from www.esri.com.
[14] Naoum, S., & Tsanis, I. K. (2004). RANKING SPATIAL
INTERPOLATION TECHNIQUES USING A GIS-BASED DSS.
Global Nest: the Int. J, 1 (6), 1-20. (NAOUM & TSANIS, 2004).
[15] Erdogan, S. (2009). A comparison of interpolation method for producing
digital elevation models at the field scale. EARTH SURFACE
PROCESSES AND LANDFORMS, 34 (3), 366-376.
[16] Zhang, X., & Srinivasan, R. (2009). GIS-BASED SPATIAL
PRECIPITATION ESTIMATION:A COMPARISON OF
GEOSTATISTICAL APPROACHES. JOURNAL OF THE AMERICAN
WATER RESOURCES ASSOCIATION, 45 (4), 894-906.
[17] Johnston, K., Hoef, J., Krivoruchko, K., & Lucas, N. (2001). Using
ArcGIS Geostatistical Analyst (pp. 33-116). United States of America.:
Esri Press.
[18] Wang, S., Huang, G. H., Lin, Q. G., Li, Z., Zhang, H., & Fana, Y. R.
(2014). Comparison of interpolation methods for estimating spatial
distribution of precipitation in Ontario, Canada. CLIMATOLOGY.
[19] Guidelines for Drinking-water Quality (THIRD EDITION ed., Vol. 1).
(2008). Geneva: WHO.
[20] Basharat M. and, Rizvi S. A. (2011, April), Groundwater Extraction
and Waste Water Disposal Regulation Is Lahore Aquifer At Stake
With As Usual Approach, Pakistan Engg. Congress; World Water DayApril 2011, P: 112-13.
[21] Akram T. and Gabriel H.F. (2007). Urban Water Cycle Management of
Lahore, Pakistan, ESDev-2007 Second International Conference on
Environmentally Sustainable Development organized by COMSATS
Institute of Information Technology, Abbottabad, Pakistan, 26-28
August 2007,ISBN 978-969-8779-13-9.
[22] Pokhrel, R. M., Kuwano, J., & Tachibana, S. (2013). A kriging method
of interpolation used to map liquefaction potential over alluvial ground.
Engineering Geology, 152 (1), 26-37.
[23] Ninyerola, M., Pons, X., & Roure, J. M. (2007). Objective air
temperature mapping for the Iberian Peninsula using spatial
interpolation and GIS. International Journal Of Climatology, 27 (9),
12311242.
[24] .Milillo, T. M., & Gardella Jr, J. A. (2006).Spatial statistics and
interpolation methods for TOF SIMS imaging. Applied surface science,
252(19), 6883-6890.
[25] Teegavarapu, R. S., Meskele, T., & Pathak, C. S. (2012). Geo-spatial
grid-based transformations of precipitation estimates using spatial
interpolation methods. Computers&Geosciences, 40, 28-39.
[26] Ashiq, M. W., Zhao, C., Ni, J., & Akhtar, M. (2010). GIS-based highresolution spatial interpolation of precipitation in mountainplain areas
of Upper Pakistan for regional climate change impact studies.
Theoretical and Applied Climatology, 99 (3-4), 239-253.
[27] Zawadzki, J., Cieszewski, C. J., & Lowe, R. C. (2005). Applying
Geostatistics for Investigations of Forest Ecosystems Using Remote
Sensing Imagery. Silva Fennica, 39 (4), 599-617.
[28] Li, j., & Heap, A. D. (2008). A Review of Spatial Interpolation Methods
for Environmental Scientists. Australia: Commonwealth of Australia.
[29] Srivastava, R. M. (2013). Geostatistics: A toolkit for data analysis,
spatial prediction and risk management in the coal industry.
International Journal of Coal Geology, 112 (1), 2-13.
[30] Andrews, B. D., Gares, P. A., & Colby, J. D. (2002). Techniques for GIS
modeling of coastal dunes. Geomorphology, 48 (1), 289-308.
[31] Hanna, K., Clark, D., & Slocombe, D. (2008). Transforming Parks and
Protected Areas: Policy and Governance in a Changing World (p. 68).
New Tork: Routledge.
[32] Omran, E.-S. E. (2012). Improving the Prediction Auccuracy of Soil
Mapping through Geostatistics. International Journal of Geosciences, 3,
574-590.
[33] Wu, Y.-H., Hung, M.-C., & Patton, J. (2013). Assessment and
visualization of spatial interpolation of soil pH values in farmland.
Precision Agriculture, 14 (6), 565-585.
[34] Chung, J.-w., & Rogers, J. D. (2012). Interpolations of Groundwater
Table Elevation in Dissected Uplands. Ground Water, 50 (4), 598-607.
[35] Siska, P. P., & Hung, I.-K. (2001). Assessment of Kriging Accuracy in
the GIS Environment. the 21st Annual Esri International User
conference. San Diego.
[36] Chang, k.-t. (2010). INTRODUCTION TO GEOGRAPHIC
INFORMATION SYSTEM. New York, America: McGraw-Hill.
332
[37] Goovaerts, P. (2000). Geostatistical approaches for incorporating

elevation into the spatial interpolation of rainfall. Journal of hydrology,
228(1-2), 113-129.
[38] Trochu, F. (1993). A contouring program based on dual kriging
interpolation. 9 (3), 160-177.
[39] Varouchakis, E. A., & Hristopulos, D. T. (2013). Comparison of
stochastic and deterministic methods for mapping groundwater level
spatial variability in sparsely monitored basins. Environmental
Monitoring and Assessment, 185 (1), 1-19.
333
Proceedings
Proceedings
Thermal Design and Analysis of PNSS-1 Satellite

M.T.Hussain, A.Abbas, B.Batul, M.U.Sadiq
Satellite Research and Development Center
SUPARCO
Lahore, Pakistan
Email: engrtassawer@hotmail.com, beenish.batul@gmail.com
AbstractPakistan National Student Satellite (PNSS-1) is a first
of its own class microsatellite being developed by Pakistan Space
and Upper Atmosphere Research Commission (SUPARCO) in
collaboration with academia. PNSS-1, weighing 50 kg, is a 3-axes
stabilized microsatellite that is designed to operate in 650 km Sun
Synchronous orbit. The efficient thermal control of a satellite is
indispensable as it maintains a specified temperature range for
the proper functioning of all the equipment on-board. In PNSS-1,
a passive thermal control is used because it is simple, lightweight, reliable and consumes no power in comparison to that of
the active thermal control. The incident heat loads from the
external environment, including Sun, Albedo, and Earth IR, were
taken into account for thermal design. Subsequently, various
operating modes were investigated to identify the expected worst
hot and cold cases that the satellite would be experiencing during
its mission life. Thermal analysis of PNSS-1 satellite was
performed using Thermal Desktop Software. The predicted
temperatures were then used to specify the size of radiator panel
to provide optimal thermal environment. A mathematical model
was also developed to simulate the orbit environment and to
predict the temperatures for the worst conditions. The results
from the mathematical model were compared with those from
the Thermal Desktop and found be in good agreement. The final
predicted temperatures fulfill the thermal control requirements
of on-board equipment to ensure their efficient performance.
Different parameters including thermal contact conductance,
insulation thickness, and optical properties of materials were also
evaluated to select the optimum thermal hardware.
KeywordsMicrosatellite,
Synchronous, Radiator
Thermal
Control,
Sun
I. INTRODUCTION
A number of programs are proposed and successfully
employed in different parts of the world in which universities
are directly engaged in different projects for the design,
manufacturing and launching of small satellites. These
projects include:
a. National Space Grant Student Satellite Program by
NASA
b. Student Space Exploration and Technology Initiative
(SSPETI) Express Sat. by ESA
c. American Student Moon Orbiter (ASMO) by
NASA
d. European Student Earth Orbiter (ESEO) by ESA
e. CubeSat Programs initiated by University of Cal
Poly, University of Arizona, TU Delft, University of
Tokyo, etc.
Student satellite is the emerging dimension of space
research at the moment. It serves many purposes including
K. Hayat
Department of Mechanical Engineering
The University of Lahore
Lahore, Pakistan
low cost Commercial Off The Shelf (COTS) component based
research, design, technology validation, and most importantly
human resource development.
Over the last two decades, there has been a considerable
increase in the use of satellite based technologies such as
vehicle navigation using Global Positioning System (GPS),
long distance phone calls, Digital Video Broadcasting,
Satellite Phone and weather forecast etc.
All over the world, public/private sector and academia
collaboration brings advancements and breakthrough in the
technological dimensions. However, in Pakistan, this type of
collaboration is rare. Joint ventures with industry and
universities are essential to bring Pakistan at par with other
nations of the world in the field of science and technology. It
will not only produce cost-effective solutions for the
challenging problems but will also contribute towards
professional grooming of the future workforce of the country.
In this regard SUPARCO has taken initiative in the form of
Pakistan National Student Satellite Program (PNSSP). The
proposed program is a way forward for a sustainable and
progressive student satellite development under the umbrella
of SUPARCO. This program will provide the platform for
collaborative efforts in real world space engineering
applications for the academia.
PNSS-1 is a three axis stabilized 50 kg class microsatellite
and is planned to be launched in a circular Sun Synchronous
orbit having an altitude of 650 km. It has three payloads
including high resolution narrow swath colored imaging
camera, wide swath colored imaging camera and scientific
experiment. The mission life will be 1 year from the launch
vehicle ignition till the satellites declared End-of-Life (EOL).
PNSS-1 will have dimensions of 470 mm x 450 mm x 470
mm.
Large satellites have inherited advantage of large thermal
mass and power storage capability which increases the heat
carrying and power storage capacity of the satellite. This large
thermal mass helps to maintain isothermal condition during
the eclipse period of the satellite and power storage capability
ensures the availability of power for heaters.
Microsatellites face rapid temperature swings due to small
thermal mass and shortage of power due to small power
storage capability making the thermal design of small
satellites a challenging task. Its necessary to maintain the
allowable temperature range of all onboard units under low
thermal mass and power shortage conditions. PNSS-1 thermal
control is primarily passive and provides the required thermal
environment to the onboard units throughout the mission life
of the satellite.
334
Similarly the values used for worst cold case are [2]:
II. ORBITAL CONDITIONS

For the mission to succeed, the PNSS-1 must survive the
rigors of the space environment from its launch till the end of
life. The extreme environments encountered by the PNSS-1
throughout the whole mission drive the thermal design of the
satellite. In orbit, PNSS-1 is exposed to three major heat loads;
direct solar heating from Sun; reflected Sun radiation from
Earth atmosphere called Earth albedo and IR radiation emitted
by earth due to its temperature.
The external heat radiations and beta angle () define the
worst hot and cold scenario encountered in the orbit. The
angle of an orbit is the minimum angle measured from the
solar vector to the orbit plane. Maximum eclipse duration
occurs for low angles and minimum or no eclipse conditions
occur at higher angles. The angle varies over a small range
during the year due to change in declination of Sun from
+23.45 to -23.45 deg. Variation of angle for PNSS-1 over a
year is shown in Fig. 1. For = 26 which occurs during
winter solstice, the satellite spends more time in the Sun, as a
result the orbit average solar load increases. On the other hand
for = 17, which occurs during summer solstice, the orbit
average solar radiation decreases. Due to seasonal variations
and degradation of thermo-optical properties of the radiator
area, the satellite faces worst hot and worst cold scenarios. For
about 94 min circular orbit at an altitude of 650 km, the values
used for the worst hot case are [2]:
Proceedings
=17
Earth Albedo Coefficient = 0.24
Earth IR Intensity = 218 W/m2
Solar Flux = 1322W/m2
Internal Heatload = 29.8W
III. PNSS-1 CONSTRUCTION
PNSS-1 is the first satellite of PNSS program having a

weight of 50 kg and launch compatibility with multiple
launchers as piggyback or cluster payload as well as with the
SUPARCOs indigenous Satellite Launch Vehicle (SLV) as
dedicated piggyback. The primary structure of PNSS-1 is a
box with dimensions of 470 mm (Z) x 470 mm (X) x 450 mm
(Y). It is subdivided into three modules; Payload Module,
Housekeeping Module, and ACS (Attitude Control
Subsystem) Module as shown in the exploded view in Fig. 2.
All modules are made of 3 mm thick Aluminum alloy plates
and reinforced with ribs, which are assembled together with
fasteners. Solar panels are mounted on four sides +X, -X, -Y
and Z of satellite (body fixed solar panels). Solar panels
dont block the environmental heat loads effectively so no heat
= 26
Earth Albedo Coefficient = 0.31
Earth IR Intensity = 244 W/m2
Solar Flux = 1414W/m2
Internal Heatload = 46.3W
Fig. 2. Exploded View of PNSS-1
dissipating unit is directly mounted on the inside of the solar

panel. Almost all the units are placed on bases of the ACS,
Payload and housekeeping module.
The configuration of the satellite is as follows:
Fig. 1. Annual Beta angle variation
335
Battery, Data Handling Unit (DHU), Power

Distribution Unit (PDU) and Power Control Unit
(PCU) are placed in Housekeeping Module as
shown in Fig. 3.
Proceedings
IV. THERMAL CONTROL SYSTEM OF PNSS-1

A. Determination of Radiator Area
Space batteries operate at low temperatures in comparison
to other subsystem units. Therefore, this requirement is a
major factor in calculation of radiator area. The +Y panel
receives minimum external flux from the environment.
Therefore, radiating window is placed on this panel. All the
high heat dissipating units are mounted directly on the
radiator. Battery is also placed here due to its narrow
temperature range. To make the initial approximation of
required radiator area and heater power, following energy
balance equation is used [3].
Fig. 3. Housekeeping Module
Onboard Computer (OBC), GPS, Reaction Wheels, Star

Tracker, Magnetometers, Rate Sensors and Magnetorquer
Rods are placed on ACS Module as shown in Fig. 4.
Fig. 4. ACS Module
PPU, Payload Transmitter, Narrow View Camera, Nanocamera, Experimental Payload, Telecommand (TC)
Receiver, Telemetry (TM) Transmitter are placed in
Payload Module as shown in Fig. 5.
Fig. 5. Payload Module
One Magnetorquer Rod, Sun sensors and Antennas

are placed outside the panels.Triple Junction Gallium
Arsenide (GaAs) Solar cells with anti-reflective coating are
placed on four panels of spacecraft. Lithium-Ion Battery is
used to store and supply power.
(1)
Where is the emissivity of the spacecraft radiator surface,
is the Stefan-Boltzmann constant (5.67x10-8 W/m2-K4), Arad
is the radiator surface area (m), Ts is the average temperature
of the spacecraft (K), As is the surface area (m), Fs,e is the view
factor between the spacecraft and the Earth, I EIR is the
intensity of Earth IR, is the surface solar absorptivity,
is
the area perpendicular to the Sun (m), and Isun is the solar heat
flux (W/m2), is the Earth albedo coefficient, Fs,se is the view
factor between the spacecraft and sunlit Earth, and Q internal is
the internal heat generation (W).
B. Thermal Design of PNSS-1
PNSS-1 is small satellite with limited mass and power
budget. Keeping in view the above constraints in thermal
design of satellite, no active technique is used. The thermal
control of satellite is achieved by adjusting the radiator area
and proper selection of thermal finishes. Passive thermal
control system in conjunction with space proven hardware is
used to provide reliable thermal control for satellite. During
the initial layout, some units were mounted on the inside of
body mounted solar panels but it was found that the
temperature fluctuation of those units was very large. So, after
having discussion with structure subsystem team, those units
were placed on the base of the module. Similarly modules
were isolated from the harness tray with access panel having
cutouts. But it was found that its necessary to remove the
cutouts and thus this change was also implemented.
Adaptor deck (+Y) remains in shadow throughout the
orbit that is why this panel is selected for radiator area.
Radiator area is covered with white paint because it has high
IR emissivity and low solar Absorptivity. The mission life of
the satellite is one year and +Y panel is not directly exposed to
Sun so darkening effect of white paint does not pose any
problem. Also, white paint is very cheap in comparison with
optical solar reflector. The portion of the panel which is not
painted white is covered with 15 layer Multilayer Insulation
(MLI) blanket. MLI blanket have alternative layers of
Aluminized Mylar and non metallic Dacron mesh enclosed in
thick Kapton outer and inner cover.
The interior surfaces of House-keeping module except
336
the area under the unit are covered with white paint. The
external surfaces of all units are painted with Chemglaze Z306
black paint for radiative heat exchange. All units have wet
contact of 1000 W/m2K with the base of the module.
ACS module lies in the middle of payload and Housekeeping module and has the major ACS units. To minimize
the temperature gradient, all the units are painted with black
paint. All the internal module surfaces are covered with white
paint. Payload module which lays on top of other modules and
having the major payload units has the similar thermal finishes
as the other two modules.
The list of materials and finishes used and their respective
thermophysical and optical properties are given in Tables I
and II respectively.
TABLEI
Density
Specific
Heat
Effective
Emissivity
W/m.K
kg/m3
J/kg.K
e*
Aluminum
120
2700
900
Solar cells
100
5774
700
MLI
1740
1004
0.03
FR4
0.343
1850
880
Kapton Tape
0.12
1420
1.09
TABLE II
Name
module. All the external faces of spacecraft's panels are placed

in outer radiation analysis group which is considered facing
the space node at 4K temperature and receives environmental
heat loads.
THERMOPHYSICAL PROPERTIES
Conductivity
Name
Proceedings
Fig. 6. Wireframe View of PNSS-1 Thermal Model
All the Aluminum panels and the units are modeled as

diffusion nodes. MLI blanket is modeled as Arithmetic nodes
above Aluminum panels with effective emissivity value,
which will be used to generate radiation conductor between
panel and MLI blanket. The body mounted solar cells are
modeled in three layers, representing FR4, Kapton film and
Aluminum. The wireframe view of the PNSS-1 Thermal
model is shown in Fig. 6.
THERMAL OPTICAL PROPERTIES

Solar Absorption
B. Thermal Analysis
Two orbit configurations are analyzed to represent the worst
hot and cold cases.
Case1, which represents hot operating case conditions,
include maximum heat dissipation mode, maximum solar
radiation at the highest angle, maximum earth infrared (IR)
and albedo radiation and maximum solar absorptance on the
external surface that is end of life properties of thermal
hardware.
Emittance
BOL
EOL
White Paint
0.19
0.19
0.88
MLI
0.45
0.50
0.67
Black Paint
0.95
0.85
Solar cells
0.91
0.91
0.81
V. THERMAL MODELING AND ANALYSIS OF PNSS-1

The transient energy balance (2), used to calculate the
temperature distribution of the satellite thermal mathematical
model is as follows:
(2)
Where Qsun, QEarth, QAlbedo, are external Heat loads, QDissipation
is internal Heatload and mcp is the thermal mass. Rij and Cij
are radiation and conduction heat exchange factors between
the nodes.
A. Thermal Modeling
The detailed thermal model (TM) is developed in Thermal
Desktop Software. The model consists of 1952 TD/RC nodes
and 233 non graphical insulation nodes. Each unit is modeled
as a single isothermal node. Each module is modeled as a
separate radiation analysis group and is radiatively decoupled
from other radiation analysis groups. Within each module,
radiation is exchanged among the units and inside walls of the
337
Fig. 7. Temperature Distribution in Hot Case
Proceedings
for both the worst cases. The results are shown in Fig. 7 to 10
and are tabulated in table III.
TABLE III
PREDICTED TEMPERATURES
Cold Case (C )
Unit
Min
Battery
Data Handling Unit
PCU
PDU
OBC
GPS
GPS
Reaction Wheel1
Reaction Wheel 2
Reaction Wheel 3
Star tracker
Magnetometer1
Magnetometer2
Rate Sensors1
Rate Sensor2
Magnetorquer Rod
PPU
Payload Transmitter
Narrow View Camera
Nano Camera
Experimental P/L
TC Receiver
TM Transmitter
Magnetorquer Rod
Sun Sensor
Solar Panel
Fig. 8. Temperature Variations in Hot Case
Case2, which represents cold operating case conditions,

includes minimum heat dissipation mode, minimum solar
loads at the lowest angle, minimum Earth IR and albedo
radiation, and minimum solar absorptance and beginning of
life properties of thermal hardware.
13.6
10
9
10
15.5
9
15
14
15
17
15.5
7
15
13.5
14
16
8
14
15
14
15
12
18
14
13.5
-25
Max
17.5
14
12
14
21.5
31
21.5
20
16
18.6
20
32
20
19.5
19.5
17.5
30
20
21
17
24
33
22
20
16.5
55
Hot Case (C )
Min
23
17
24
21
25
18
26
26.5
25
28
25
26.5
16
23.5
24.5
26.5
19
28
30
27
28
22
30
25.5
25
-15
Max
27.4
19
29
26
31
39
33
28
31
29.5
30
32
43
30
30
28.5
41
35
38
31
38
43
36
32
28.5
60
Design
Range
(C)
1 +30
+45
+45
+45
+45
-1
+50
-1
+50
-40 +70
-40 +70
-40 +70
-40 +70
-40 +70
-40 +70
-30 +50
-30 +50
-50 +85
0 +45
+45
+45
+45
+45
+45
+45
-50 +85
+85
-1
+ 100
Comparison of predicted temperatures in the worst hot and

cold case shows the temperature of units is maintained within
the required design range.
VI. PARAMETRIC STUDY
Fig. 9. Temperature Distribution in Cold Case
Fig. 10. Temperature Variations in Cold Case
The transient analysis is carried out for 23 orbital periods
All the units of spacecraft operate optimally in a certain

temperature range. Therefore there is a need to optimize
thermal design and assess the results. As design selected is
purely passive, we are not left with any option other than
optimal selection of thermal finishes and contact conductance
to meet the temperature requirements of units. To maintain the
temperature range of onboard units, parametric study was
conducted. The parameters studied are effective emissivity of
MLI, contact conductance, angle and emissivity of black and
white paints respectively. Variation of parameters and their
effect on the final temperature is listed in the Table V. The
plus and minus symbols in Table V represent the increase and
decrease in temperature from the nominal value respectively.
It can be observed that for all the parameters except angle,
decrease in property value from the nominal results in increase
of temperature. Similarly, decrease in temperature is observed
when the property value is increased from the nominal value,
except for the angle case. For angle, there is direct
relationship between change in temperature and value. Of all
the listed parameters, beta angle has major effect on the
temperature of the units.
338
TABLE V
Property Name
MLI Effective
Emissivity
Contact Conductance
(W/m2C)
Angle
0.03
1000
26
Black Paint Emissivity

White Paint Emissivity
significant effect on the temperature of satellite units as

compared to contact conductance and emissivity of surfaces.
This thermal design will be further verified by thermal
testing as the program will proceed. The results will be
correlated with the predicted temperatures. Any discrepancy
will be incorporated in subsequent thermal model to ensure
accuracy of thermal control of satellite for the successful
mission.
PARAMETRIC STUDY
Nominal
0.85
0.85
Modified
0.01
0.05
800
1200
22
30
0.75
0.95
0.75
0.95
Proceedings
Change in
Temperature
(0C)
+0.9
-0.9
+0.2
-0.2
-7
+6
+0.12
-0.05
+0.06
-0.09
ACKNOWLEDGMENT
The authors are thankful to the colleagues of GSPFE
division at Satellite Research and Development Center Lahore
for their timely help and support. In particular, we appreciate
the support provided by Dr. Hafiz Muhammad Saleem Iqbal
who gave suggestions to improve the work.
VII. CONCLUSION
Preliminary Design Review (PDR) of the PNSS-1 has been
successfully held. Preliminary thermal design of PNSS-1 has
been presented in the PDR meeting. The passive thermal
design is selected to meet the temperature requirements of the
onboard units. The design has been verified by analysis
performed using Thermal Desktop Software in integration
with RadCAD, a radiation analyzer. The predicted
temperatures are in agreement with the thermal requirements.
The results of parametric study show that orbit angle has
REFERENCES
[1]
[2]
[3]
339
Karam, R. D., Satellite Thermal Control for System Engineer,

Progress in Astronautics and Aeronautics Series, AIAA, New York,
1998, pp. 106
D.G. Gilmore, Spacecraft Thermal Control Handbook, The Aerospace
Corporation, 2002, pp. 26 - 29.
Andrew D. Williams, Scott E. Palo, Issues and Implications of the
Thermal Control Systems on the Six Day Spacecraft, 4th Responsive
Conference, Los Angeles, CA, 2006.
Proceedings
Hybrid Median-Nulling Scheme for Impulsive Noise

Mitigation in OFDM Transmission
Zehra Ali
Satellite Research and Development Center,
Space and Upper Atmosphere Research Commission,Karachi, Pakistan
ali_zehra@hotmail.com
AbstractOrthogonal Frequency Division Multiplexing
(OFDM) is a rapidly growing technology in digital
communications. OFDM performance is degraded by impulsive
noise which can be mitigated by introducing different nonlinearity schemes at the OFDM receiver like clipping, nulling,
hybrid clipping-nulling and replacement. Clipping and nulling
are the simplest schemes. Hybrid clipping-nulling combines the
advantages of both clipping and nulling and performs better than
the other schemes. Replacement scheme defines the clipping level
equal to the average of the transmitted signal and gives
comparable results as hybrid clipping-nulling. This article
reviews some impulsive noise mitigation schemes suggested in
literature and proposes a new scheme, known as hybrid mediannulling, which clips the signal values at the level defined by
median of the received signal samples. Simulation results show
that the proposed scheme results in better signal to noise ratio
and symbol error rate as compared to other schemes.
KeywordsOFDM; impulsive noise; impulsive noise mitigation
schemes; threshold
I.
INTRODUCTION
Orthogonal Frequency Division Multiplexing is a digital

modulation technique and is employed in various wired and
wireless standards [1], such as European Digital Terrestrial
Video Broadcasting (DVB-T), 4G mobile satellite systems,
Global Navigation Satellite Systems (GNSS) and in power
line communication (PLC) modems [2]. Despite of many
advantages of OFDM system, random occurrence of high
power impulsive noise [1] severely degrades OFDM system
performance especially in DVB-T systems and ultimately
results in lower signal to noise ratio (SNR) and higher symbol
error rate (SER) [3].
It is found that the impulsive noise has higher spectral
density than background noise [4]. Different methods to model
impulsive noise (IN) effects have been reported in literature
[5]-[7]. The most common and widely accepted model of
impulsive noise is Middleton Class A model [8]. Therefore,
we also modelled the impulsive noise in our work using
Middleton Class A model. This is further explained in section
II.
Several techniques to reduce the effect of impulsive noise
(IN) in OFDM transmission system have been proposed.
Applying clipping and/or blanking non-linearity to the
received signal samples before they pass through a
conventional OFDM demodulator are among the simplest
approaches to mitigate impulsive noise at the OFDM receiver
340
[9]. These approaches basically involve one or both of the two

transformations to be applied at the samples affected by
impulsive noise.
The first transformation is to limit the maximum amplitude
of the received signal samples up to a certain fixed level. This
is called clipping. The second transformation is to multiply the
samples affected by impulsive noise by zero. This is called
nulling or blanking because it completely removes the signal
and replaces it with zero. Both nulling and clipping decisions
are governed by the threshold which is usually fixed at the
OFDM receiver [10]. It is the threshold level which decides if
the sample is affected by impulsive noise or not. Since nulling
is an extreme transformation, the threshold for nulling is
usually kept higher than threshold for clipping when using a
combination of both schemes, called as, hybrid clippingnulling [11].
Hybrid clipping-nulling is suitable for both high and low
probabilities of impulsive noise [9]. In conventional clipping,
the clipping level or the maximum level of output samples is
kept equal to the clipping threshold. In most of the cases, the
clipping threshold is a fixed value unless any adaptive
threshold or threshold optimizing technique block is used in
the OFDM transmission system [12]. In [11], a simpler IN
mitigation scheme is proposed, known as, replacement
scheme. In replacement scheme, instead of a fixed threshold
for clipping, the clipping level is replaced by mean or average
of the original transmitted samples which are noiseless. The
results of hybrid clipping-nulling and replacement schemes are
comparable. However, replacement scheme is simpler because
it involves only two conditional actions. Since it is not feasible
to practically find the mean of original signal at the receiver,
therefore, theoretical value of mean is used.
In this paper we introduce a new hybrid median-nulling
scheme in which the clipping threshold level is based upon the
median of the received signal. Since median is less sensitive to
the outliers, it gives a normal value of the skewed distribution
[13]-[14]. Because of the addition of high amplitude impulsive
noise, the practical mean value of the received samples
changes drastically while median gives the true middle value
of the whole signal received. This is why, hybrid mediannulling gives better resulting signal to noise ratio (SNR) and
symbol error rate (SER) as compared to hybrid clippingnulling and replacement scheme.
This paper is organized as follows. OFDM System Model
and Impulsive Noise (IN) model are described in section II. In

section III, we discuss different IN mitigation schemes and
explain the proposed hybrid median-nulling scheme.
Simulation results are shown in section IV. Finally, the
conclusion is mentioned in section V.
Proceedings
is the complexzero mean white Gaussian noise with variance
i2 (1 / 2) E[ g n ].
2
The transmitted signal is also passed through an additive

transmission channel, including AWGN (Additive White
Gaussian Noise) denoted by wn and has variance
w2 (1 / 2) E[ wn ].
2
II.
OFDM SYSTEM MODEL
In a conventional OFDM transceiver, the original signal to

be transmitted is considered noiseless. The data is divided into
N subcarriers which are orthogonal to each other [15]. N data
symbols carried in vector x k, from 16-QAM (Quadrature
Amplitude Modulation) scheme, are transformed into time
domain signal x nthrough IDFT (Inverse Discrete Fourier
Transform) as shown in (1).
xn
1
N
N 1
X
k 0
e j 2nk / N
Thus, the noisy channel can be characterized by the signal-toAWGN ratio; SNR 10 log 10 (1 / w2 ) and signal-to-IN ratio;
SINR 10 log 10 (1 / i2 ) . . The received time-domain signal
can be expressed as ( rn )
if bn 0
x wn
rn n
xn wn in if bn 1
(1)
III.
Where N is the number

(1) of subcarriers, j 1 , x k and
xn are of same length.
The variance of x n is given by
2
2 (1 / 2) E x
1 and 1 / N is the normalizing factor
for IDFT.
A. Impulsive Noise Model
Impulsive Noise is classified into two classes; Middleton
Class A and Class B [5]. In DVB-T applications, where there
is a high risk of impulsive noise occurrence, Middleton Class
A type is considered [16]-[17], with a probability density
function (PDF) as shown in (2).
e m Am
1
p( z )
e
m0
m!
2 m2
z2
2 2
m
Clipping
rn
y n j arg(
rn )
Te
| rn | T
| rn | T
c.
(7)
Hybrid clipping-nulling
rn
y n T1 e j arg(rn )
0
(3)
(6)
Nulling
r | rn | T
yn n
0 | rn | T
(2)
Where A is the average number of impulses, 2 is the total

noise power, subscript G denotes Gaussian noise power and I
denotes impulsive (non-Gaussian) noise power.
For simplification, we consider a special case of Middleton
Class A noise model [18] in which IN is denoted by in and is
modelled as Bernoulli-Gaussian random process as follows:
in bn g n
a.
b.
where

2
2
2
A
, 2 G2 I2 , G2
m
1
I
(3)
IN MITIGATION SCHEMES
In this section, we review clipping, nulling, hybrid clippingnulling [9] and replacement [11] schemes and propose a new
IN mitigation scheme named as hybrid median-nulling.
Let y n be the signal sample obtained after applying any IN
mitigation scheme to rn and Tis the threshold level. The
impulsive noise mitigation schemes are given as follows:
(5)
| rn | T1
T1 | rn | T2
| rn | T2
(8)
Where T2>T1, optimized value of T2 is 1.4T1 [9].

d.
Replacement
rn
yn
j arg( rn )
| x | e
| rn | T
| rn | T
(9)
(4)
Where bn is the Bernoulli process of sequence bn 1

or bn 0 with probability of p or1 p respectively, and g n
341
Where x is the average or mean of transmitted samples.

Theoretical values of x are mentioned in [11].
E[ x n ]
SNR
E[ y x 2 ]
n
n
e. Proposed Scheme: Hybrid median-nulling
rn
y n | m | e j arg(rn )
| rn | T1
T1 | rn | T2
| rn | T2
(10)
Where m median(rn ) . If N samples of r is arranged

in ascending order such that:
r1 min( r ), r2 , r3 ,..., rn1 , rN max( r )
, the statistical
median (m ) of (r ) is defined as in [13]
r N 1
2
m
1
rN r N
2 2 1 2
N
0
2
N
0
2
(12)
Fig. 2. shows that the SNR achieved by median-nulling

scheme is significantly higher than the SNR obtained from
replacement and clipping- nulling scheme
(11)
It is worthwhile mentioning here that the signal to noise

ratio (SNR) and symbol error rate (SER) of all the above
schemes largely depend upon the threshold level (T).
Optimum threshold is the threshold value which gives
maximum SNR and minimum SER varies for each scheme
[18]. We consider varying values of threshold in simulation in
section IV.
IV.
Proceedings
Fig 2 (a) p=0.01
SIMULATION RESULTS AND DISCUSSIONS
The computer simulation of the OFDM system is carried

out with 16-QAM modulation, signal to noise ratio;
SNR=40dB, signal to impulsive noise ratio; SINR=-15dB and
number of subcarriers, N = 10^5. Probability(p) of IN
occurrence is considered to be p=0.01, 0.03, and 0.1, which
implies that 1%, 3%, and 10% of the received samples will be
affected by IN respectively. Therefore, with 10^5 subcarriers,
the average number of IN pulses received within each OFDM
symbol is (pN), i.e., about 1000, 3000, and 10000 IN pulses
per OFDM symbol for p=0.01, 0.03, and p=0.1, respectively.
The simulation is carried out with threshold levels varying
from 0 to 15 with increment of 0.1. The simulation block
diagram of OFDM system simulated is shown in Fig. 1.
Fig 2 (b) p=0.03
Fig 2 (c) p=0.1

Fig. 2. The SNR of IN mitigation schemes with different threshold values
SNR = 40dB and SINR=-15dB.
This relative gain in SNR of proposed scheme as compared

to other schemes is further illustrated in Fig. 3. where it is
calculated by using (13) obtained from [18] in an AWGN and
IN channel.
Fig. 1.Simulation block
The output SNR is obtained by using (12) as mentioned in

[9].
342
SNRA
SNR _ Gain(dB) 10 log 10

SNR
B
Proceedings
(13)
Where SNRA denotes the SNR obtained by hybrid mediannulling scheme and SNRB denotes the SNR obtained by hybrid
clipping-nulling or replacement schemes.
Fig. 4 (a) p=0.01
Fig. 3 (a)
Fig. 4 (b) p=0.03
Fig. 3 (b)
Fig. 3. The relative gain of IN mitigation schemes with different threshold

values when p=0.01, 0.03 and 0.1. SNR = 40dB and SINR=-15dB. (a)
Relative gain with respect to hybrid clipping nulling scheme. (b) Relative
gain with respect to replacement scheme
Fig. 4 (c) p=0.1

Fig. 4. The SER of IN mitigation schemes with different threshold values
when SNR = 40dB and SINR=-15dB.
System performance of the proposed scheme is compared

with replacement and hybrid clipping-nulling schemes in
terms of Symbol Error Rate (SER) also by using (14) as
mentioned in [19]. SER comparison is shown in Fig. 4. Fig. 4.
proves that the SER of hybrid median-nulling is considerably
lower than the other schemes.
SER
3
erfc SNR
2
(14)
343
Table I and II summarize the performance comparison and

prove that the proposed scheme outperforms the other
schemes in different scenarios of impulsive noise occurrence
in terms of both signal to noise ratio (SNR) and symbol error
rate (SER).
TABLE1. PERFORMANCE COMPARISON OF HYBRID CLIPPING-NULLING,

REPLACEMENT AND HYBRID MEDIAN-NULLING SCHEME IN TERMS OF SNR.
p
Hybrid clippingNulling Scheme
References
[1]
Maximum SNR (dB)

Replacement
Hybrid Median Scheme
Nulling Scheme
[2]
0.01
12
12
17
0.03
7.9
7.5
13
[3]
0.1
0.1
2.4
7.5
[4]
TABLE II. PERFORMANCE COMPARISON OF HYBRID CLIPPING-NULLING,

REPLACEMENT AND HYBRID MEDIAN-NULLING SCHEME IN TERMS OF SER
p
Hybrid clippingNulling Scheme
Maximum SER
Replacement
Scheme
[5]
Hybrid Median Nulling Scheme
[6]
[7]
0.01
0.1
0.1
0.001
0.03
0.4
0.4
0.06
[8]
0.1
0.8
0.83
0.4
[9]
V.
CONCLUSION
[10]
We introduce a new IN mitigation scheme, called as, hybrid

median-nulling scheme. The proposed scheme is similar to
hybrid clipping-nulling scheme in terms of conditions
depending upon the two thresholds; T1 and T2. However, in
hybrid median-nulling, the maximum level of output samples
is defined by the median of the received signal. Therefore, the
proposed scheme differs from hybrid clipping-nulling in terms
of the clipping level which is not equal to the fixed threshold
value and varies according to the OFDM signal characteristics.
Hybrid median-nulling also outperforms the replacement
scheme which uses mean or average value of the noiseless
samples as the clipping level. The simulation results show that
the proposed scheme results in better performance than the
other schemes at different probabilities, provided that the
threshold values T1 and T2 are optimized.
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
Mengi, A., & Vinck, A. H.: Successive impulsive noise suppression in

OFDM. Proc. IEEE Int. Symp. Power Line Communications (ISPLC),
Rio de Janeiro, Brazil, March 2010, pp. 33-37
Suraweera, H. A., Chai, C., Shentu, J., & Armstrong, J.:Analysis of
impulse noise mitigation techniques for digital television systems. Proc.
8th Int. OFDM Workshop (InOWo03), Hamburg, Germany, Sep. 2003,
pp. 172-176
Armstrong, J., and Suraweera, H.A.: Impulse noise mitigation for
OFDM using decision directed noise estimation. Proc. IEEE ISSSTA
2004, Sydney, Australia, Aug./Sep. 2004, pp. 174-178
Cuntic, P. &Bazant, A.:Analysis of modulation methods for data
communications over the low-voltage grid. Proc. IEEE Int. Conf.
Telecommun., Zagreb, Croatia, Jun. 2003, 2, pp. 643648.
Middleton, D.: Canonical and quasi-canonical probability models of
class A interference, IEEE Trans. Electromagn. Compat, May 1983,25,
(2), pp. 76106
Middleton, D.: Non-gaussian noise models in signal processing for
telecommunications: New methods an results for class A and class B
noise models, IEEETrans.Inf.Theory, May 1999, 45, (4), pp.11291149
Sanchez. M.G., Haro, L. de., Ramon, M.C., Mansilla, A. Ortega, C.M. &
Oliver, D. : Impulsivenoisemeasurementsandcharacterization in a UHF
digital TV channel, IEEE Trans. Electromagn. Compat., May 1999, 41,
(2), pp. 124136.
Middleton,
D.:
Statisticalphysicalmodelsofelectromagneticinterference,
IEEE
Trans.
Electromagn. Compat., Aug. 1977,EMC-19, (3), pp. 106127
Zhidkov, Sergey V.: Analysis and comparison of several simple
impulsive noise mitigation schemes for OFDM receivers, IEEE
Transactions on Communications, Jan. 2008, 56, (1), pp. 5-9
Zhidkov, Sergey V.: Impulsive noise suppression in OFDM based
communication systems, IEEE Transactions on Consumer Electronics,
Nov. 2003, 49, (4), pp. 944-948
Papilaya, V. N., & Vinck, A. H.:Investigation on a new combined
impulsive noise mitigation scheme for OFDM transmission. 17th IEEE
International Symposium on Power Line Communications and Its
Applications (ISPLC), Johannesburg, South Africa, Mar. 2427, 2013,
pp. 8691
Epple, U., & Schnell, M.:Adaptive threshold optimization for a
blanking nonlinearity in OFDM receivers. Global Communications
Conference (GLOBECOM) IEEE, California, USA, December 2012, pp.
3661-3666
Weisstein EW, Statistical Median. From MathWorld--A Wolfram Web
Resource,
http://mathworld.wolfram.com/StatisticalMedian.html,
accessed 5 February 2015
Craig, P., Kennedy, J., & Cumming, A.: Animated interval scatter-plot
views for the exploratory analysis of large-scale microarray time-course
data, Information Visualization, 2005, 4, (3), pp. 149-163
Debbah, M.: Short introduction to OFDM, White Paper, Mobile
Communications Group, Institut Eurecom, February 2004, pp. 0-1-0-11
Ndo, G., Siohan, P., &Hamon, M., Adaptive noise mitigation in
impulsive environment: Application to power-line communications.
IEEETrans.PowerDel., Apr. 2010, 25, (2), pp.647656
Amirshahi, P., Navidpour, S., &Kavehrad, M.: Performanceanalysis of
uncoded and coded OFDM broadband transmission over low voltage
power-line channels with impulsive noise, IEEE Trans. Power Del.,
Oct. 2006, 21, (4), pp. 19271934
Alsusa, E., &Rabie, K.M.: Dynamic peak-based threshold estimation
method for mitigating impulsive noise in power-line communication
systems,IEEE Trans. on Power line, Oct. 2013,28, (4), pp. 2201-2208
Barry, J.R., Lee, E.A., and Messerschmitt, D.G.: Digital Communication'
(Springer, 3rd edn. 2003)
This article is sponsored by Pakistan Space and Upper Atmosphere Research Commission
344
Proceedings
Proceedings
Measurement & Testing Techniques of

Performance Parameters for Electric Servo
Actuators
Naveed Riaz
Faisal Rehman
Mechatronics Engineering Department

CEME, NUST
engrnaveedriaz@gmail.com
Mechatronics Engineering Department

CEME, NUST
fasial_superb@yahoo.com
actuator, and is required to be kept as minimum as possible. It

is the band in the output at during which the input changes but
the output remains unchanged. Dead band was calculated by
generating command voltage and observing the output
displacement, a delay was observed that was calculated in
open loop at bi-directional operation of actuator.
Backlash, an important non linear parameter, is present
inherently in the mechanical systems, and was calculated
using both the power and no load condition of actuator by
locking the output and applying a step input.
Abstract - Servo Actuators are found in various dynamic

systems like CNC machines & robotics, and are designed
on the basis of desired functional parameters like Output
Torque, Slew rate and rated Power. Parameters like
Friction and Mechanical Backlash offers negative impact
on the functionality of such systems. This paper discusses
the measurement and testing techniques of these
parameters. Torque & slew rate will be calculated
theoretically (mathematically), and then the results will be
verified by applying a testing technique known as torsion
bar technique. This paper also discusses simple techniques
for calculating Mechanical backlash & Friction/Dead
band and suggests ways for eliminating their effects.
II. TORQUE MEASUREMENT USING TORSION BAR

Maximum Torque of servo actuator can only be calculated
at maximum power. Specifications of servo actuator under
test are mentioned below:
Index terms - Servo Actuator, Torque, Slew rate, Mechanical

Backlash, Dead band, Torsion Bar.
I. INTRODUCTION
Servo Actuators are the type of actuators that work in
closed loop i.e. they take feedback from their outputs for
further decisions. They are mostly found in position and
velocity control of various machines. There are many types of
servo actuators e.g. hydraulic actuators, pneumatic actuators,
electric actuators etc. This paper discusses electric servo
actuators and especially rotary actuators. Performance
parameters of rotary servo actuators are proposed and testing
techniques are devised based on close simulation to actual
system applications. Output torque and slew rate, friction &
mechanical backlash are the basic parameters which are kept
in mind while designing any type of servo actuator.
When a mechanical member is loaded with a moment
about a longitudinal axis, they are said to be in torsion, and
the applied moment is then termed as torque. Torsion in a bar
directly simulates to the amount of applied torque provided
the material properties of the bar kept constant. This property
of bar helps to calculate the required torque, by directly
measuring, the amount of twist produced. For this purpose, a
bar was calibrated and a jig was designed specifically to
measure the twist by directly applying the load on a moment
arm of known length. Torsion jig was adjusted to calibrate bar
at required torque value. Slew rate is the rate at which servo
actuator achieves the desired position at rated torque. It was
calculated by measuring the rise time of output. Friction or
dead band is the undesirable parameter in the design of servo
Power = 26 Watts, Operating Voltage = 24 Volts, Max.

Current limit = Amps, Max. Output Torque = 5.25 N-m
(Theoretical or Mathematical).
Consider a straight hexagonal steel bar of length l. If we
twist this bar around the axis resisted by the bar's torsion
resistance, the effective bar spring rate can be determined by
the diameter, length and material of the bar. The bar twists
about its longitudinal axis and its free end deflects through an
angle .
The angular deflection due to applied torque is;
Where;
l
JG
(1)
= Twist Angle (in radians).

T = Torque applied (in Nm).
l = length of the bar (in meters).
J = polar area moment of inertia of hexagonal
bar.
G = Rigidity Modulus.
345
Eq. (1) is the basis for measuring the torque of servo

actuator. The advantage of designing a jig is that we need not
to calculate J and G of hex bar. These factors are calculated
by calibration of bar.
Torsion bar is calibrated in order to calculate the scale

factor i.e. the amount of torque required to twist the bar to
1degree and to check any permanent deformation in the bar
and its spring rate. Spring rate, is the torsional resistance of
the bar. The amount of twist angle in torsion bar is
calculated by using Eq. (1) assuming, Torque = 1.5 Nm,
Length of bar = 305mm and specific J and G of selected
material of bar.
J is calculated by dividing hexagon bar into 6 triangles
then calculating the polar moment of inertia of each triangle
and adds it up.
The polar moment of inertia of hex bar is found to be 15.75
mm^4 and G = 80.8 GPa. G is the modulus of rigidity of
steel.
The twist angle is;
= (1.5Nm * 0.120m) / (1.575E -11 * 8.08E10)
(2)
= 0.1414 radians = 8.10
Slope, m = (V1-V0) / (t1-t0)

(5)
Where V0 and V1 are the initial and final voltages and t0 and
t1 is the initial and final time taken to achieve V1
respectively.
V1
V0
t0
Calibration is performed on a loading jig. The loading jig

comprises of a torsion bar clamped with the wheel of jig
assembly. The other end of torsion bar is fixed. The amount
of twist produced in the torsion bar while the other end is
fixed simulates the loading on the bar. The amount of the
twist is measured using potentiometer. When the output gear
of the servo actuator is clamped with the wheel of the loading
jig using link rod, the twist in the torsion bar simulates the
loading on the output gear.
Now we try to achieve that angle i.e. 8.1 by adjusting
the rear end plate of the torsion jig backward or forward ( to
change the effective length of the torsion bar) until a reading
of 0.451 0.05V i.e. equivalent to 8.1 0.8 is obtained. We
have locked the plate at this particular position and calculated
the torsion bar scale factor i.e.
8.10 twist is generated by torsion bar at = 1.5 Nm torque.
1 twist is generated by torsion bar at = 1.5/8.1 = 0.185
Nm torque.
So torsion bar scale factor k= 0.185Nm/
(3)
t1
= slope or slew rate
Fig.1. Measurement of slew rate
IV. MEASUREMENT OF MECHANICAL BACKLASH

Backlash is any undesirable play in the axis due to
looseness \ looseness in mechanical parts. When command is
given to move the axis, the motor turns shortly before
movement is initiated. This delay is backlash.
Mechanical backlash is the amount of play between
meshing gears. It causes difference b/w input and output of
gears i.e.
If A is the input revolution of driving gear and B is the
output revolution of driven gear, then for perfect gears;
A = Bx
Servo actuator is operated at its maximum power. The

amount of twist produced due to maximum force applied by
actuator is measured by potentiometer; putting it into
following equation, yields the maximum torque;
T = k
Proceedings
(6)
Where, x is the gear ratio.

But in actual scenario;
A = (B-y) x. Where, y is the amount of backlash or play
between the gears.
(4)
Output gear of actuator was locked at zero position

feedback potentiometer gives zero deflection. Supply voltage
and current is set to almost half the actual values required to
drive the actuator, say if actuator operates at 24V & 1A.
Apply 24V and 0.5 Amps. It is necessary because servo is
presently at locked rotor position so it will draw maximum
current at once and that could damage the control circuitry
When system was powered the feedback potentiometer gives
the deflection even the out put is locked, this deflection is the
backlash b/w gears. Depending upon the placement of
feedback potentiometer either on output gear or at some
intermediate position where backlash y is calculated, the net
backlash of servo is calculated by dividing y with the x the
gear ratio of servo actuator.
III. MEASUREMENT OF SLEW RATE

Slew rate is basically a constant speed that actuator
maintains at any load. This speed depends upon the
proportional gain you set in your controller to drive servo
actuator.
Slew rate calculation is performed on same jig setup by
connecting actuator with torsion bar clamped at one end. In
this experiment oscilloscope is used to measure the deflection
of torsion bar. Step input of actuator generates a slope of
deflection recorded on the oscilloscope. Servo actuator moves
to its end limits. The output step response is rate at which the
deflection signals or output arm reaches the desired end limit.
We calculate the rate by following formula;
346
Net backlash = y/x
(7)
V. MEASUREMENT OF FRICTION/DEADBAND
Proceedings
REFERENCES
[1] Larry B. Li, System and method for controlling high side slew
rate in a spindle motor driver, Texas Instruments Incorporated.
USA, Patent No. US6072289 A
[2] Asao Watanabe, Shuji Satoh, System and method for
controlling high side slew rate in a spindle motor driver, Nf.
T&M. Systems. Inc., Nippon Pulse Motor Co., Ltd. USA,
Patent No. US5424960 A
[3] E. M. H. Kamerbeek "On the Theoretical and Experimental
Determination of the Torque in Electrical Machines", Philips
Tech. Rev., no. Supplement, No. 4, 1970
[4] K. Kemper , D. Koepl and J. Hurst "Optimal passive dynamics
for torque/force control", Int. Conf. Robot. Autom., 2010
[5] Wilson, Rich. "Lexus, the car company: this all new sports
sedan is just the beginning of things to come for Lexus.",
Automotive Industries, March 2005.
Dead band or friction of servo actuator is the minimum

amount of power consumed to set the output arm of servo in
continuous motion. When voltage starts to rise from zero
there is no deflection of output gear up to certain level of
voltage. Initially when voltage begins to rise, the torque
generated is used to overcome the friction of actuator, keep on
increasing voltage at certain instance output gear starts
deflecting but still its motion is not uniform because there is
still some dynamic friction present in the system. When
output gear starts to move uniformly at that point the friction
is overcome and the voltage is noted. This voltage range
defines the dead band region of servo actuator. This value is
very helpful in setting the proportional gain of controller. One
rule is the minimum proportional gain of controller should be
the end voltage of dead band region. Greater the dead band
more is the friction so more power controller has to provide to
the actuator to overcome friction.
ACKNOWLEDGMENT
The authors would like to give special thanks to their
professors and colleagues for their valuable comments and
suggestions in respect of this work.
347
Proceedings
IEEE Submitted Papers

1
Saad Arif
Track Fusion of Legacy Radars using Weighted Averages
Dr. Muhammad Yasir
Design & Development of Air Bearing Table for On Ground Testing of AOCS
Hardware
M. Hammad Riaz
Single Event Effect Testing of Commercial Off-The-Shelf Components
Khalid Saeed
Crack Growth Performance of Aluminum Plates Repaired with Composite and

Metallic Patches under Fatigue Loading
M. Naveed Akhtar
Steady State Heat Transfer using Galerkin Finite Element Method
Adil Murtaza
Role of magnetic anisotropy on Magnetostrictive P roperties of quasi-binary

erromagnetic Tb1-xGdxFe2 system
Adil Murtaza
Morphotropic phase boundary in pseudo-binary Ferromagnetic Tb1-xGdxFe2 System
Hamza Shams
Evaluation of Commercial Oleophobic Coatings for Aerospace Applications in Harsh

Environments
Osama Arshed Malik
Impact Resistance Analysis Using Multiply Fabric Orientations
10
Shaheryar Atta Khan
Evaluation of Additive Manufacturing Techniques for Fabrication of Propellers for

SUAVs
11
Nasar A. Mubarak
Feasibility To Adapt Modifications In The Extant Turbojet Engine Test Bed For The
Ground Test Run Of Turbofan Engine
12
Abir Raza Baig
Fingerprint Enhancement over the Past Few Years
13
Adil Nawaz
Performance Analysis of Supervised Image Classification Techniques for the

Classification of Multispectral Satellite Imagery
14
Dr. M Yousaf Hamza
Evolution Behavior of Hyperbolic Secant Pulse in Normal and Anomalous Dispersion

Regimes of Single Mode Optical Fiber Communication
15
Muhammad Ahmad
CCDF of EVM for Digital Modulation Schemes over Fixed Satellite Service
16
Raheel Muzzammel
Analysis of Path Losses and Isolation in 60 GHz Networks and Design of Intelligent
Neighbor Scanning
17
Wasim Nawaz
Improvement of Gain in Dual Fed X Band Isoflux Choke Horn Antenna for use in
LEO satellite mission
18
Zehra Ali
Hybrid Median-Nulling Scheme for Impulsive Noise Mitigation in OFDM

Transmission
19
Jabir Shabbir Malik
Remote Sensing of Ocean, Ice and Land Surfaces Using Bistatically Reflected
GNSS Signals From Low Earth Orbit
20
Maria Mahmood
Improving the detection and estimation processes of acquisition module in Global

Navigation Satellite Systems (GNSS)
21
Muhammad Fayyaz
The Trends/ Variations of Ionospheric Parameters (hmF2, foF2) between Observatory

and International Reference Ionosphere Web Model Values.
22
Wajeeha Najeeb
Implementation And Evaluation of Vehicle Tracking System
23
Maria Neelum
Study of Foliar Rust Diseases of Wheat Crop in Chakwal District Using GIS tools
24
Muhammad Shahzad Anjum
Nonlinear synchrotron self-Compton modelling of blazars
25
Muhammad Touseef Ikram
Do We Really Have to Consider Data Mining Techniques for Meteorological Data
26
Sana Liaquat
Object Detection and Depth Estimation of Real World Objects using Single Camera
27
Gulnaz Ahmed
Analyzing Algorithms in Wireless Body Area Sensor Networks: A Survey
348

Conference Proceedings ICASE 2015 FINAL

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Conference Proceedings ICASE 2015 FINAL

Uploaded by

Copyright:

Available Formats

Fourth International Conference on Aerospace Science & Engineering (ICASE 2015)

Fourth International Conference on

Institute of Space Technology

Fourth International Conference on Aerospace Science & Engineering (ICASE 2015)

Fourth International Conference on Aerospace Science & Engineering (ICASE 2015)

ICASE ORGANIZING COMMITTEE

INTERNATIONAL SCIENTIFIC COMMITTEE

NATIONAL SCIENTIFIC COMMITTEE

Fourth International Conference on Aerospace Science & Engineering (ICASE 2015)

(Secretary ICASE 2015)

Mr. Zeeshan Fareed

(Marketing and Publicity)

Mr. Raza Butt

(Media and Public Relations)

Mr. Waqas Jilani Joiya

(Logistics and Operations)

Mr. Muhammad Adeel

Mr. Waqas Ramzan

Fourth International Conference on Aerospace Science & Engineering (ICASE 2015)

Fourth International Conference on Aerospace Science & Engineering (ICASE 2015)

Dr. Najam Abbas Naqvi

Fourth International Conference on Aerospace Science & Engineering (ICASE 2015)

A Comparison Study of Advanced state Observers for Quad rotor UAV

Fuzzy Temperature Controller for Induction Heating

Qazi Ejaz ur Rehman

Stability and Control Solution of Quad-Copters

Detection of Fire Hotspots dealt by Emergency First Responders in

Prospects of Airborne Wind Energy Systems in Pakistan

Integrated use of Potential Rainwater Harvesting Site for Agriculture

Muhammad Usman Saleem

Urban change detection of Lahore (Pakistan) using the Thematic

Improving physical and mechanical properties of medium density fiber

Finite Element Analysis of Tool Wear in Ultrasonically Assisted

Engr Numan Khan

Finite Element Simulation of Composite Body Armor

Malik Abid Hussain

GIS for estimating Optimized Water Demand Using Sustainable Water

Farzan Javed Sheikh

Optimized Threshold Calculation based on Received Signal

Muhammad Altaf Khan

Intelligent Detection of Distributed Flooding Attack in Wireless Mesh

Touch panel Based Restaurent Automation Using Zigbee Technology

Introducing space plant biology to students through hands-on activities

The Political and Economic Feasibility of Current Space Resource

Feasibility Study to Install Fire Fighting Equipment on a Cargo

Najam ud Din Ahmad

Syed Jahanzeb Hussain Pirzada

Optimize Manufacturing of unidirectional carbon prepregs for space

Design of a Fuzzy Logic Water Level Controller

The use of Nuclear reactor in Space Applications:Propulsion and Power

Fourth International Conference on Aerospace Science & Engineering (ICASE 2015)

Impact of Thermal Aging on Microstructure and Mechanical Properties

Madni Shifa Ullah Khan

Effect of aryl diazonium salt functionlization on the electrical properties

Design of C band Slotted Waveguide Array antenna with high

Muhammad Nauman Hussain

Risk Areas Mapping and Identification of Hotspots on the RoadNetwork of Lahore

Analysis of Recent Drought Based on NDVI and Meteorological Data

Ali Jan Hassan

Assessment of Urban Growth of Karachi: From A Tiny Town to A Meta

Abdur Raqeeb Gaziani

JUPITER, The Gas GIANT

Shakeel Ahmad Waqas

M Ameer Umar Malik