Professional Documents
Culture Documents
WSN 3
WSN 3
Course
Outline Radio Technology Primer
Week-1
Available Wireless Technologies
Over the past many years, the world has become increasingly mobile. As a result, traditional
ways of networking the world have proven inadequate to meet the challenges posed by our
new collective lifestyle. If users must be connected to a network by physical cables, their
movement is dramatically reduced. Wireless connectivity, however, poses no such restriction
and allows a great deal more free movement on the part of the network user.
SENSING AND
SENSORS
Sensors link the physical with the digital world by capturing and revealing real-world
phenomena and converting these into a form that can be processed, stored, and acted
upon. Integrated into numerous devices, machines, and environments, sensors provide a
tremendous societal benefit.
They can help to avoid catastrophic infrastructure failures, conserve precious natural
resources, increase productivity, enhance security, and enable new applications such as
context-aware systems and smart home technologies.
The phenomenal advances in technologies such as very large scale integration (VLSI),
microelectromechanical systems (MEMS), and wireless communications further contribute
to the widespread use of distributed sensor systems.
SENSING (DATA
ACQUISITION)
SENSORS
CLASSIFICATION
Which sensors should be chosen for an application depends on the physical property to be
monitored, for example, such properties include temperature, pressure, light, or humidity.
Besides physical properties, the classification of sensors can be based on a variety of other
methods, for example, whether they require an external power supply. If the sensors
require external power, they are referred to as active sensors. That is, they must emit some
kind of energy (e.g., microwaves, light, sound) to trigger a response or to detect a change in
the energy of the transmitted signal.
On the other hand, passive sensors detect energy in the environment and derive their
power from this energy input – for example, passive infrared (PIR) sensors measure
infrared light radiating from objects in the proximity.
SENSORS
CLASSIFICATION
The classification of sensors can also be based on the methods they apply and the electrical
phenomena they utilize to convert physical properties into electrical signals.
Resistive sensors rely on changes to a conductor’s electrical resistivity, ρ, based on physical
properties such as temperature.
Capacitive sensors, which can be used to measure motion, proximity, acceleration, pressure,
electric fields, chemical compositions, and liquid depth.
Inductive sensors are based on the electrical principle of inductance, that is, where an
electromagnetic force is induced by a fluctuating current. Inductance is determined by the
dimensions of the sensor (cross-sectional area, length of coil), the number of turns of the
coil, and the permeability of the core. Changes in any of these parameters (e.g., caused by
movements of the core within the coil) change the inductance. Inductive sensors are often
used to measure proximity, position, force, pressure, temperature, and acceleration.
SENSORS
CLASSIFICATION
Piezoelectric sensors use the piezoelectric effect of some materials (e.g., crystals and
certain ceramics) to measure pressure, force, strain, and acceleration. When a pressure is
applied to such a material, it causes a mechanical deformation and a displacement of
charges, proportional to the amount of pressure.
The main advantage of piezoelectric devices over other approaches is that the piezoelectric
effect is not sensitive to electromagnetic fields or radiation.
SENSORS
CLASSIFICATION
Type Examples
Temperature Thermistors, thermocouples
Two sensor fields monitoring two different geographic regions and connecting to the Internet using their base stations.
WIRELESS SENSOR
NETWORK
The capabilities of sensor nodes in a WSN can vary widely, that is, simple sensor nodes may
monitor a single physical phenomenon, while more complex devices may combine many
different sensing techniques (e.g., acoustic, optical, magnetic).
They can also differ in their communication capabilities, for example, using ultrasound,
infrared, or radio frequency technologies with varying data rates and latencies. While simple
sensors may only collect and communicate information about the observed environment,
more powerful devices (i.e., devices with large processing, energy, and storage capacities)
may also perform extensive processing and aggregation functions.
WIRELESS SENSOR
NETWORK
Such devices often assume additional responsibilities in a WSN, for example, they may form
communication backbones that can be used by other resource-constrained sensor devices to
reach the base station.
Finally, some devices may have access to additional supporting technologies, for example,
Global Positioning System (GPS) receivers, allowing them to accurately determine their
position. However, such systems often consume too much energy to be feasible for low-cost
and low-power sensor nodes.
History of
WSN
DARPA:
Distributed Sensor Nets Workshop (1978)
Distributed Sensor Networks (DSN) program (early 1980s)
Sensor Information Technology (SensIT) program
UCLA and Rockwell Science Center
Wireless Integrated Network Sensors (WINS)
Low Power Wireless Integrated Microsensor (LWIM) (1996)
UC-Berkeley Smart Dust project (1999)
Concept of “motes”: extremely small sensor nodes
Berkeley Wireless Research Centre (BWRC)
PicoRadio project (2000)
MIT
μAMPS (micro-Adaptive Multidomain Power-aware Sensors) (2005)
WSN Communication
Characteristics of typical WSN:
low data rates (comparable to dial-up modems).
energy-constrained sensors.
IEEE 802.15.4 is an example for a protocol that has been designed specifically for short-range
communications in WSNs.
low data rates.
low power consumption.
widely used in academic and commercial WSN solutions.
Single-Hop versus Multi-
Hop
Star topology:
Every sensor communicates directly (single-hop) with the base station.
May require large transmit powers and may be infeasible in large geographic areas.
Mesh topology:
Sensors serve as relays (forwarders) for other sensor nodes (multi-hop).
May reduce power consumption and allows for larger coverage.
Introduces the problem of routing.
Single-Hop versus Multi-
Hop
CHALLENGES :
ENERGY
Sensors typically powered through batteries
Replace battery when depleted.
Recharge battery, e.g., using solar power.
Discard sensor node when battery depleted.
For batteries that cannot be recharged, sensor node should be able to operate
during its entire mission time or until battery can be replaced.
Energy efficiency is affected by various aspects of sensor node/network design
Physical layer:
Switching and leakage energy of CMOS-based processors.
CHALLENGES :
ENERGY
Medium access control layer:
Contention-based strategies lead to energy-costly collisions.
Problem of idle listening.
Network layer:
Responsible for finding energy-efficient routes.
Operating system:
Small memory footprint and efficient task switching.
Security:
Fast and simple algorithms for encryption, authentication, etc.
Middleware:
In-network processing of sensor data can eliminate redundant data or aggregate sensor readings.
CHALLENGES : SELF
MANAGEMENT
Ad-hoc deployment
Many sensor networks are deployed “without design”.
Sensors dropped from airplanes (battlefield assessment).
Sensors placed wherever currently needed (tracking patients in disaster zone).
Moving sensors (robot teams exploring unknown terrain).
Other terminology
Self-Organization is the ability to adapt configuration parameters based on system and
environmental state.
Self-Optimization is the ability to monitor and optimize the use of the limited system
resources.
Self-Protection is the ability recognize and protect from intrusions and attacks.
Self-Healing is the ability to discover, identify, and react to network disruptions.
CHALLENGES : WIRELESS
NETWORKING
The reliance on wireless networks and communications poses a number of challenges to a
sensor network designer. For example, attenuation limits the range of radio signals, that is, a
radio frequency (RF) signal fades (i.e., decreases in power) while it propagates through a
medium and while it passes through obstacles. The relationship between the received
power and transmitted power of an RF signal can be expressed using the inverse-square
law:
which states that the received power Pr is proportional to the inverse of the square of the
distance d from the source of the signal. If distance is doubled, power requirement increase
to 4 times.
As a consequence, an increasing distance between a sensor node and a base station rapidly
increases the required transmission power. Therefore, it is more energy-efficient to split a
large distance into several shorter distances, leading to the challenge of supporting multi-
hop communications and routing.
CHALLENGES : WIRELESS
NETWORKING
Attenuation:
limits radio range
Multi-hop communication:
Increased latency
Increased failure/error probability
Complicated by use of duty cycles ----- reduce by wakeup-on-demand or adaptive duty cycle
CHALLENGES :
DECENTRALIZATION
Centralized management (e.g., at the base station) of the network often not feasible to
due large scale of network and energy constraints.
Therefore, decentralized (or distributed) solutions often preferred, though they may
perform worse than their centralized counterparts.
Example: routing
Centralized:
BS collects information from all sensor nodes
BS establishes “optimal” routes (e.g., in terms of energy)
BS informs all sensor nodes of routes
can be expensive, especially when the topology changes frequently
Decentralized:
each sensors makes routing decisions based on limited local information
routes may be non optimal, but route establishment/management can be much cheaper
CHALLENGES : DESIGN
CONSTRAINS
Many hardware and software limitations affect the overall system design
Examples include:
Low processing speeds (to save energy)
Low storage capacities (to allow for small form factor and to save energy)
Lack of I/O components such as GPS receivers (reduce cost, size, energy)
Lack of software features such as multi-threading (reduce software complexity)
CHALLENGES : SECURITY
REFERENC
By Kazem Sohraby, Daniel Minoli, Taieb Znati
To maximize the opportunity for widespread and cost-effective deployment of WSN, one
needs to make use of existing and/or emerging Commercial Off The-Shelf (COTS) wireless
communications and infrastructures rather than having to develop an entirely new, specially
designed apparatus. WSNs can use a number of wireless COTS technologies, such as
Bluetooth/Personal Area Networks (PANs), ZigBee, wireless LANs (WLAN)/hotspots,
broadband wireless access (BWA)/WiMax, and 3G.
PAGE 35
RADIO TECHNOLOGY
PRIMER
The electromagnetic spectrum provides an unguided medium (channel) for point-to-point
and/or broadcast radio transmission.
Radio transmission is usually (frequency)-bandlimited by design. The analog bandwidth of
the channel (the slice of electromagnetic frequency domain used) determines how much
information (analog or digital) can be transmitted over the channel.
A transmission channel in general, and a radio-based channel in particular, is never perfect
because it is subjected to external (and even internal) noise sources; noise has a tendency to
degrade, disrupt, or otherwise affect the quality of an intelligence-bearing signal.
A lot of radio-transmission engineering has to do with how to deal with the noise problem;
the goal is nearly always to optimize the signal-to-noise ratio, subject to specified
constraints (e.g., bandwidth requirements, cost, reliability, power consumption, equipment
and antenna size).
PAGE 36
Propagation and Propagation Impairments
Issues of interest in radio design include, among others, propagation, impairments, environment (i.e.,
indoors–outdoors, unobstructed–obstructed, benign–hostile, etc.), sensitivity, antenna design, channel
bandwidth (analog and/or digital), and frequency of operation. Many design factors (e.g., propagation,
attenuation, impairments) are related parametrically to the frequency band in use. In particular,
directionality becomes more of an issue at higher frequency ranges; also, generally, bandwidth increases
as one moves to higher-frequency bands (given that larger portions of the spectrum are in principle
available).
The most basic model of radio-wave propagation typically found in WSN environments involves
the direct or free-space wave. PAGE 37
Propagation and Propagation Impairments
In this model, radio waves emanate from a point source of radio energy, traveling in
all directions in a straight line, filling the entire spherical volume of space with
radio energy that varies in strength with a 1/(distance)2 rule (or 20 dB per tenfold
increase in distance); attenuation in environments that are not free space (e.g.,
waters, coaxial cable, heavily wooded areas, confined rooms or structures) is
considerably more severe.
PAGE 38
Basic Phenomenon affecting signals
Reflection: When size of object is larger than the wavelength of propagating wave.
Signal strength fluctuations caused by the fact that the composite signal received comprises a number of
components from the various sources of reflections from different directions as well as scattered and/or
diffracted signal components affect both mobile and stationary receivers, whether the receivers are indoors
or outdoors.
The intrinsic electromagnetic (radio) signal strength attenuation caused by these phenomena is called a
large-scale effect; signal-strength fluctuations related with the motion of the broadcasting or receiving
antenna are called small-scale effects.
PAGE 40
Multipath
Metallic materials as well as dielectrics (or electrical insulators) cause reflections. When
multiple signal propagation paths exist, the actual signal level received is the vector sum
of all the signals incident from any direction or angle of arrival.
Some signals will aid (constructively reinforce) the direct path; others will subtract
(destructively interfere with or vector-cancel out) from the direct signal path.
Channel performance varies with user location and time, and the radio propagation pattern
is complex. One needs to deal with multipath scattering from nearby objects, shadowing
from dominant objects, and attenuation effects from various physical phenomena. All of
these factors result in rapid fluctuations of received power; even when the device mobile
is stationary, the signals received may fade, due to movement of surrounding objects.
PAGE 41
Outdoor Radio
Propagation
Figure describes pictorially issues related to outdoor propagation. For indoor propagation applications,
the signal decays much faster: walls, floors, and furniture attenuate or scatter radio signals; also, the
coverage is restricted to the local environment by walls and the like. The path loss formula is :
PAGE 42
Path loss
PAGE 43
Attenuation contribution factors
Additional contributing factors include the following:
People moving around (additional multipath-induced attenuation of up to 10 dB).
Buildings with few metal and hard partitions: root-mean-square (rms) delay spread
of 30 to 60 ns (equaling several Mbps without equalization).
Buildings with metal or open aisles: rms delay spread of up to 300 ns (hundreds of
kbps without equalization).
Between floors: Concrete or steel flooring yields less attenuation than that of steel
plate flooring Metallic-tinted windows yield greater attenuation, 15 dB for first-floor
separation, 6 to 10 dB for the next four floors, 1 to 2 dB for each additional floor of
separation.
PAGE 44
Attenuation contribution factors
The indoor signal strength received depends on the office plan, construction materials, density of
personnel, furniture, and so on (e.g., wall losses, 10 to 15 dB; floor losses, 12 to 27 dB; delay spread, varies
between 15 and 100 ns, requiring sophisticated equalization techniques to achieve acceptable bit-error
rates).
Table depicts signal attenuation values for signals typically used in networking and telecom applications.
A drawback of higher-frequency bands (e.g., 5 GHz for IEEE 802.11a applications) compared to lower-
frequency bands (e.g., 2.4 GHz for IEEE 802.11b/g applications) is the shorter wavelength of the signal at
the higher band.
It turns out that short-wavelength signals have more difficulty propagating through physical obstructions
encountered in an office (walls, floors, and furniture) than do those at longer wavelengths.
PAGE 45
Attenuation contribution factors
Error bursts are an outcome of fades in radio channels. Doppler-induced frequency or
phase shifts due to motion can also cause loss of synchronization. Errors increase as the
bit period approaches the delay spread.
The typical acceptable BER for data communications is 106.
Strategies for overcoming errors include antenna diversity, forward error correction
techniques, and traditional automatic repeat request (retransmission protocol for blocks
in error).
The outdoor-to-indoor penetration or building loss depends on building materials,
orientation, layout, height, percentage of windows, and transmission frequency.
The strength of the signal received increases with increasing building height; the
penetration loss decreases with increasing frequency (e.g., 6 dB loss through windows).
PAGE 46
Effects in Industrial and Factory
environment
In an industrial environment, care is needed when placing sensors in order to minimize
interference. One needs to keep WNs away from other sources of radio-frequency
interference (RFI), such as brush-type electrical motors, other radio transmitters or
transceivers, or unshielded computer equipment and/or cables.
Sensors that must be located near such devices should connect to the transceiver via a
short piece of shielded cable so that they can stay as far away as possible from the
source of the RFI.
In a factory environment, large iron and steel structures may create multipath
problems. As noted, multipath propagation occurs when nearby metal reflects the radio
signal in the same way that a mirror reflects light. The receiver detects multiple signals
simultaneously—the original and the reflections—and cannot decode any of them.
Moving the receiving or transmitting antenna just a few inches is sometimes enough to
fix this problem. PAGE 47
Multipath Mitigation
Techniques
Radio system design: redundant paths for each receiver, if possible.
Antenna system design: dual diversity antennas used at each receiver.
Signal/waveform design: spread-spectrum radio design with the highest feasible chip rate.
Building/environment design: not much can be done in this area unless RF-friendly
greenfield buildings are constructed.
PAGE 48
Use of the industrial, scientific and medical (ISM) band (at 2.4 GHz— more exactly,
2.412 to 2.484 GHz), and of the Unlicensed Network Information Infrastructure (U-
NII) band (at 5 GHz) does not require a license.
However, there still are technical guidelines that must be followed in terms of the
radiated power, radiation pattern, and so on.
PAGE 49
Modulation
Modulation is the overlay of an intelligent signal over an underlying carrying signal, which
is then transmitted over the medium in question (be it a cable, wireless, or fiber-optic
medium).
Baseband applications are those applications where the coded signal is carried directly
over a medium without having to overlay it onto a carrier signal.
Non-baseband systems use modulation; baseband systems do not. In traditional
environments modulation allows transmission over long distances (e.g., tens to hundreds
of miles); baseband systems usually are limited to the carriage of information over a
fraction of a mile.
Traditional wired LAN systems are baseband systems: The signal is encoded by some
appropriate mechanism (e.g., Manchester encoding) and then transmitted over
unshielded twisted-pair cable.
Analog radio and TV transmission use modulation.
PAGE 50
Modulation
Types
Modulation
PCM
PAM PWM PPM
AM FM PM ASK FSK PSK QAM
PAGE 51
AVAILABLE WIRELESS
TECHNOLOGIES
As we noted, two frequency bands are typically used by WNs: the ISM band and the
U-NII band.
Indoor and outdoor interference arises from both natural sources and/or phenomena
(e.g., loss or attenuation, absorption, fading, multipath) as well as from other users in
proximity utilizing these ‘‘unprotected bands.’’ A WSN will experience interference
whether it uses one of the IEEE PAN/LAN/MAN technologies or even some other
generic radio technology.
PAGE 52
AVAILABLE WIRELESS
TECHNOLOGIES
IEEE PAN/LAN/MAN technologies are broadly implemented technologies and are probably the
ones utilized in the majority of (commercial) WSNs on a going-forward basis.
Protocols determine the physical encoding of signal transmitted as well as the data link layer
framing of the information.
Channel-sharing and data- and event-handling procedures are also specified by the protocol.
There are several wireless protocols; the most widely used are:
IEEE 802.15.1 (also known as Bluetooth)
IEEE 802.11a/b/g/n series of wireless LANs
IEEE 802.15.4 (ZigBee)
MAN-scope IEEE 802.16 (also known as WiMax)
Radio-frequency identification (RFID) tagging.
Each standard possesses different benefits and limitations.
PAGE 53
Time Synchronization
PAGE 54
Clock and the Synchronization Problem
In distributed systems, each node has its own clock and its own notion of time.
However, a common time scale among sensor nodes is important to identify causal
relationships between events in the physical world, to support the elimination of
redundant sensor data, and to generally facilitate sensor network operation. Since
each node in a sensor network operates independently and relies on its own clock,
the clock readings of different sensor nodes will also differ.
In addition to these random differences (phase shifts), the gap between clocks of
different sensors will further increase due to the varying drift rates of oscillators.
Therefore, time (or clock) synchronization is required to ensure that sensing times
can be compared in a meaningful way.
PAGE 55
Computer clocks based on hardware oscillators are essential components of all
computing devices. A typical clock consists of a quartz-stabilized oscillator and a
counter that is decremented with every oscillation of the quartz crystal. Whenever
the counter value reaches 0, it is reset to its original value and an interrupt is
generated.
Each interrupt, or clock tick , increments a software clock (another counter), which
can be read and used by applications using a suitable application programming
interface (API).
Therefore, a software clock provides a local time for a sensor node, where C(t)
indicates the clock reading at some real time t . The time resolution is the distance
between two increments (ticks) of the software clock.
PAGE 56
Comparing the local times of two nodes, the clock offset indicates the difference
between the times. Synchronization is required to adjust the time of one or both of
these clocks such that their readings match. The clock rate indicates the frequency at
which a clock progresses and the clock skew is the difference in the frequencies of two
clocks.
Perfect clocks have a clock rate dC/dt = 1 at all times, but various parameters affect the
actual clock rate, for example, the temperature and humidity of the environment, the
supply voltage, and the age of the quartz. This deviation results in a drift rate, which
expresses the rate by which two clocks can drift apart, that is, dC/dt − 1.
The maximum drift rate of a clock is expressed as ρ with typical values for quartz-based
clocks being 1 ppm to 100 ppm (1 ppm = 10−6). This number is given by the
manufacturer of the oscillator and guarantees that 1 − ρ ≤ dC/dt ≤ 1 + ρ.
PAGE 57
Figure 9.1 shows how the drift rate affects the clock reading with respect to real time, resulting in
either a perfect, fast, or slow clock.
PAGE 58
This drift rate is responsible for inconsistencies in sensors’ clock readings even after
clocks have been synchronized, making it necessary to repeat the synchronization
process periodically. Assuming identical clocks, any two clocks that are synchronized
can drift from each other at a rate of at most 2ρmax.
To limit the relative offset to δ seconds, the resynchronization interval τsync must
meet the requirement:
τsync ≤ δ/2ρmax
C(t) must be piecewise continuous, that is, a strictly monotone function of time.
PAGE 59
Types of
Synchronization
External synchronization means that the clocks of all nodes are synchronized with an
external source of time (or reference clock ). The external reference clock is an
accurate real-time standard such as Coordinated Universal Time (UTC). External
synchronization ensures both synchronization with an external source and consistency
among all clocks within the network.
Internal synchronization means that the clocks of all nodes are synchronized with each
other, without the support of an external reference clock. The goal of internal
synchronization is to obtain a consistent view of time across all nodes in the network,
even though this time may be different from any external reference times. When
nodes in a network are internally synchronized, the precision indicates the maximum
offset between any two clocks in the network.
PAGE 60
Why Time Synchronization in
WSNs?
Sensors in WSNs monitor objects and events in the physical world
Accurate temporal correlation is crucial to answer questions such as
how many moving objects have been detected?!
what is the direction of the moving object?!
what is the speed of the moving object?!
If real-time ordering of events is t1<t2<t3, then sensor times should reflect this ordering:
C1(t1) <C2(t2) <C3(t3)
PAGE 61
Why Time Synchronization in
WSNs?
Time difference between sensor time stamps should correspond to real-time
differences:
Δ = C2(t2)-C1(t1) = t2-t1
important for data fusion (aggregation of data from multiple sensors)
Synchronization needed by variety of applications and algorithms
Communication protocols (at-most-once message delivery)
Security (limit use of keys, detect replay attacks)
Data consistency (caches, replicated data)
Concurrency control (atomicity and mutual exclusion)
Medium access control (accurate timing of channel access)
Duty cycling (know when to sleep or wake up)
Localization (based on techniques such as time-of-flight measurements)
PAGE 62
Challenges for Time
Synchronization
Traditional protocols (e.g., NTP) are designed for wired networks
WSNs pose a variety of additional challenges
Environmental effects
sensors often placed in harsh environments
fluctuations in temperature, pressure, humidity
Energy constraints
finite power sources (batteries)
time synchronization solutions should be energy-efficient
Wireless medium and mobility
throughput variations, error rates, radio interferences, asymmetric links
topology changes, density changes
node failure (battery depletion)
Other challenges
limited processor speeds or memory (lightweight algorithms)
cost and size of synchronization hardware (GPS) PAGE 63
Basics of Time
Synchronization
PAGE 64
Synchronization Messages
One-Way Message Exchange
Two-way Message Exchange
Receiver–Receiver Synchronization
PAGE 65
Time Synchronization
Protocols
Reference Broadcasts Using Global Sources of Time
Lightweight Tree-Based Synchronization (LTBS)
Timing-sync Protocol for Sensor Networks
Flooding Time Synchronization Protocol
Reference-Broadcast Synchronization
Time-Diffusion Synchronization Protocol
Mini-Sync and Tiny-Sync
PAGE 66
Reference
Broadcasts
Global Positioning System (GPS) is a well-known global source of time.
Terrestrial radio stations
WWV/WWVH & WWVB (National Institute of Standards & Technology).
continuously broadcast time based on atomic clocks
Problems with these techniques:
not universally available (underwater, indoors, outer space)
need for high-power receivers
size
cost
PAGE 67
Lightweight Tree-Based
Synchronization
LTS subdivides time synchronization into two building blocks:
• A pair-wise synchronization protocol synchronizes two
neighboring nodes.
• To keep all nodes or the set of interesting nodes
synchronized to a common reference, a spanning tree from
the reference node to all nodes is constructed.
PAGE 68
Lightweight Tree-Based
Synchronization
In order to maximize the synchronization accuracy, the depth of the tree should be
minimized. This is due to the fact that the errors resulting from the pairwise
synchronizations are additive and therefore increase along the branches of the tree as
a function of the number of hops.
In LTS, a tree construction algorithm such as breadth first search is executed each time
the synchronization algorithm is executed. Once the tree has been established, the
reference node initiates the synchronization by performing the pairwise
synchronization with each of its children.
Once synchronized, each child repeats this step with its own children until all nodes of
the tree have been synchronized.
PAGE 69
Lightweight Tree-Based
Synchronization
Two different variants are proposed:
Centralized multihop LTS: The reference node – for example, a node with a GPS receiver or another
high-quality time reference – constructs a spanning tree T and starts synchronization: First the
reference node synchronizes with its children in T , then the children with their children, and so forth.
Hence, each node must know its children. There are several algorithms available for distributed
construction of a spanning tree; for LTS, two specific ones are namely the Distributed Depth-First
Search (DDFS) and the Echo algorithms.
Distributed multihop LTS: The second variant is the distributed multihop LTS protocol. No spanning
tree is constructed, but each node knows the identities of a number of reference nodes along with
suitable routes to them. It is the responsibility of the nodes to initiate resynchronization periodically.
One or more reference nodes contacted by sensors whenever synchronization is required. Nodes
determine resynchronization period based on desired clock accuracy, distance to reference node, clock
drift ρ, time of last synchronization. Node can query neighbors for pending synchronization requests,
i.e., node synchronizes with neighbor instead of reference node.
PAGE 70
Timing-sync protocol for
sensor networks (TPSN)
The Timing-sync Protocol for Sensor
Networks (TPSN) is another traditional
sender–receiver synchronization approach
that uses a tree to organize a network.
PAGE 72
Flooding Time Synchronization
Protocol
PAGE 73
Flooding Time Synchronization
Protocol
In this analysis, the wireless radio of the sensor node informs the CPU
using an interrupt at time t1 that it is ready to receive the next piece of
the message to be transmitted. After the interrupt handling time d1, the
CPU generates a time stamp at time t2. The time needed by the radio to
encode and transform the piece of the message into electromagnetic
waves is described as encoding time d2 (between t1 and t3).
The propagation delay (between t3 on node j ’s clock and t4 on node k’s
clock) is followed by the decoding time d4 (between t4 and t5). This is
the time the radio requires to decode the message from electromagnetic
waves back into binary data. The byte alignment time d5 is a delay
caused by the different byte alignments (bit offsets) of nodes j and k,
that is, the receiving radio has to determine the offset from a known
synchronization byte and then shift the incoming message accordingly.
Finally, the radio on node k issues an interrupt at time t6, which allows
the CPU to obtain a final time stamp at time t7.
PAGE 74
Flooding Time Synchronization
Protocol
Time-stamping in FTSP
sender sends single broadcast containing time stamp (estimated global time)
receiver extracts time stamp from message and time-stamps arrival (leads to global-local time
pair, providing a synchronization point)
synchronization message begins with preamble followed by SYNC bytes, data field, and CRC
preamble bytes are used to synchronize receiver radio to carrier frequency
SYNC bytes are used to calculate bit offset
PAGE 75
Flooding Time Synchronization
Protocol
Time-stamping in FTSP
multiple time stamps are used at both sender and receiver to reduce jitter of interrupt handling and
encoding/decoding times
time stamps are recorded at each byte boundary after the SYNC bytes as they are transmitted or
received!
time stamps are normalized by subtracting appropriate integer multiple of nominal byte transmission
time (e.g., approx. 417μs on Mica2)
jitter in interrupt handling can be reduced by taking the minimum of normalized time stamps
jitter in encoding/decoding can be reduced by averaging these corrected normalized time stamps
final (error-corrected) time stamp is added into data part of message
at receiver side, time stamp must further be corrected by the byte alignment time (can be determined
from transmission speed and bit offset)
PAGE 76
Reference-Broadcast
Synchronization
Key idea of RBS: in the wireless medium, broadcast messages will arrive at receivers at
approximately the same time
set of receivers synchronize with each other using a broadcast message
variability in message delay dominated by propagation delay and time needed to receive and
process incoming message (send delay and access delay are identical)
RBS critical path is short than critical path of traditional techniques
PAGE 77
Reference-Broadcast
Synchronization
The strength of RBS lies in the removal of nondeterministic synchronization errors caused by the
sender.
Example with 2 receivers:
receivers record arrival of synchronization message
receivers exchange recorded information
receivers calculate offset (difference of arrival times)
More than 2 receivers:!
maximum phase error between all receiver pairs is expressed as group dispersion
likelihood that a receiver is poorly synchronized increases with the number of receivers
(larger group dispersion)
increasing the number of broadcasts can reduce group dispersion
Offsets between two nodes can be computed as the average phase offsets for all m packets
received by receivers i and j:
PAGE 78
Reference-Broadcast
Synchronization
PAGE 79