Professional Documents
Culture Documents
Ebook Autonomous Underwater Vehicles Localization Tracking and Formation Cognitive Intelligence and Robotics Yan Online PDF All Chapter
Ebook Autonomous Underwater Vehicles Localization Tracking and Formation Cognitive Intelligence and Robotics Yan Online PDF All Chapter
Ebook Autonomous Underwater Vehicles Localization Tracking and Formation Cognitive Intelligence and Robotics Yan Online PDF All Chapter
https://ebookmeta.com/product/autonomous-underwater-vehicles-
jing-yan-xian-yang-haiyan-zhao-xiaoyuan-luo-xinping-guan/
https://ebookmeta.com/product/good-robot-bad-robot-dark-and-
creepy-sides-of-robotics-autonomous-vehicles-and-ai-1st-edition-
jo-ann-oravec/
https://ebookmeta.com/product/embedded-robotics-from-mobile-
robots-to-autonomous-vehicles-with-raspberry-pi-and-arduino-4th-
edition-thomas-braunl/
https://ebookmeta.com/product/cambridge-igcse-and-o-level-
history-workbook-2c-depth-study-the-united-states-1919-41-2nd-
edition-benjamin-harrison/
Sustainability Prospects for Autonomous Vehicles
Environmental Social and Urban 1st Edition George T.
Martin
https://ebookmeta.com/product/sustainability-prospects-for-
autonomous-vehicles-environmental-social-and-urban-1st-edition-
george-t-martin/
https://ebookmeta.com/product/received-signal-strength-based-
target-localization-and-tracking-using-wireless-sensor-networks-
satish-r-jondhale-r-maheswar-jaime-lloret/
https://ebookmeta.com/product/autonomous-and-connected-vehicles-
network-architectures-from-legacy-networks-to-automotive-
ethernet-1st-edition-dominique-paret/
https://ebookmeta.com/product/autonomous-vehicles-volume-1-using-
machine-intelligence-1st-edition-a-mary-sowjanya-syed-imran-
patel-varshali-jaiswal-imran-khan-allam-balaram-romil-rawat/
Cognitive Intelligence and Robotics
Autonomous
Underwater
Vehicles
Localization, Tracking, and Formation
Cognitive Intelligence and Robotics
Series Editors
Amit Konar, ETCE Department, Jadavpur University, Kolkata, India
Witold Pedrycz, Department of Electrical and Computer Engineering, University of
Alberta, Edmonton, AB, Canada
Cognitive Intelligence refers to the natural intelligence of humans and animals, it is
considered that the brain performs intelligent activities. While establishing a hard
boundary that distinguishes intelligent activities from others remains controversial,
most common behaviors and activities of living organisms that cannot be fully
synthesized using artificial means are regarded as intelligent. Thus the acts of
sensing and perception, understanding the environment, and voluntary control of
muscles, which can be performed by lower-level mammals, are indeed intelligent.
Besides the above, advanced mammals can perform more sophisticated cognitive
tasks, including logical reasoning, learning, recognition, and complex planning and
coordination, none of which can yet be realized artificially to the level of a baby,
and thus are regarded as cognitively intelligent.
This book series covers two important aspects of brain science. First, it attempts
to uncover the mystery behind the biological basis of cognition, with a special
emphasis on the decoding of stimulated brain signals or images. Topics in this area
include the neural basis of sensory perception, motor control, sensory-motor
coordination, and understanding the biological basis of higher-level cognition,
including memory, learning, reasoning, and complex planning. The second
objective of the series is to publish consolidated research on brain-inspired models
of learning, perception, memory, and coordination, including results that can be
realized on robots, enabling them to mimic the cognitive activities performed by
living creatures. These brain-inspired models of machine intelligence complement
the behavioral counterparts studied in traditional artificial intelligence.
The series publishes textbooks, monographs, and contributed volumes.
Autonomous Underwater
Vehicles
Localization, Tracking, and Formation
Jing Yan Xian Yang
Institute of Electrical Engineering Institute of Information Science
Yanshan University and Engineering
Qinhuangdao, Hebei, China Yanshan University
Qinhuangdao, Hebei, China
Haiyan Zhao
Institute of Electrical Engineering Xiaoyuan Luo
Yanshan University Institute of Electrical Engineering
Qinhuangdao, Hebei, China Yanshan University
Qinhuangdao, Hebei, China
Xinping Guan
Department of Automation
Shanghai Jiao Tong University
Shanghai, China
© The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature
Singapore Pte Ltd. 2021
This work is subject to copyright. All rights are solely and exclusively licensed by the Publisher, whether
the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse
of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and
transmission or information storage and retrieval, electronic adaptation, computer software, or by similar
or dissimilar methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication
does not imply, even in the absence of a specific statement, that such names are exempt from the relevant
protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this book
are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or
the editors give a warranty, expressed or implied, with respect to the material contained herein or for any
errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional
claims in published maps and institutional affiliations.
This Springer imprint is published by the registered company Springer Nature Singapore Pte Ltd.
The registered company address is: 152 Beach Road, #21-01/04 Gateway East, Singapore 189721,
Singapore
Preface
Ocean covers 70.8% of the earth’s surface, and it plays an important role in supporting
system of life on the earth. Nonetheless, more than 80% of the oceans volume remains
unmapped, unobserved, and unexplored. With regard to this, autonomous underwater
vehicles (AUVs) are emerged as a promising solution to understand and explore the
ocean, due to their autonomy and mobility. AUVs have a variety of military and
civilian applications, such as the intrusion surveillance, geographic mapping, marine
rescue, and petroleum exploration. It is worth noting that the global market for AUVs
is estimated to grow from 638 million dollars in 2020 to 1638 million dollars by 2025,
at a compound annual growth rate of 20.8%. To enable AUVs for a wider range of
application-specific missions, it is necessary to deploy multiple AUVs to cooper-
atively perform the localization, tracking, and formation tasks. However, the weak
communication of underwater acoustic communication and the model uncertainty of
AUVs make it challenging to achieve the localization, tracking, and formation tasks.
In this book, we focus on the localization, tracking, and formation issues for
AUVs, subject to weak communication in cyber-channel and model uncertainty in
physical channel. This problem is of great necessary and meaning since fundamental
guidance on the design of localization, tracking, and formation of AUVs is very
limited up to now. In this context, we devote to present the cutting-edge results
on the localization, tracking, and formation for AUVs. It highlights new research
on commonly encountered AUV system, including the ubiquitous localization of
AUVs with optimization rigid graph, the tracking for a single AUV with unknown
parameters and external disturbances, the input-constrained tracking for a single
AUV with communication delays, and the input-constrained formation of multiple
AUVs with communication delays. Moreover, this book also presents several joint
localization and tracking solutions for AUVs, including the codesign when weak
communication is considered, and the codesign when weak communication and
model uncertainty are both considered. In order to verify the effectiveness of the
proposed solution, simulation and experimental results are both provided. Finally,
future directions provide research guidance on the design for future localization,
tracking, and formation schemes on AUVs.
v
vi Preface
1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1 Underwater Observation Network . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 An Overview of the Existing Works . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2.1 Ubiquitous Localization of AUVs . . . . . . . . . . . . . . . . . . . . . . 3
1.2.2 Persistent Tracking Control for a Single AUV . . . . . . . . . . . . 9
1.2.3 Cooperative Formation for Multiple AUVs . . . . . . . . . . . . . . 15
1.3 Research Challenges and Contributions of This Book . . . . . . . . . . . . 18
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
2 Rigid Graph-Based Asynchronous Localization of AUVs . . . . . . . . . . . 25
2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
2.2 Network Model and Problem Formulation . . . . . . . . . . . . . . . . . . . . . 27
2.2.1 Network Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
2.2.2 Problem Formulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
2.3 Main Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
2.3.1 Observer-Based Motion Prediction for AUV . . . . . . . . . . . . . 31
2.3.2 Minimum Rigid Graph-Based Localization . . . . . . . . . . . . . . 35
2.3.3 Performance Analyses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
2.4 Simulation and Experimental Results . . . . . . . . . . . . . . . . . . . . . . . . . . 49
2.4.1 Simulation Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
2.4.2 Experiment Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
2.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
3 Slide Mode-Based Joint Localization and Tracking of a Single
AUV . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
3.2 System Model and Problem Formulation . . . . . . . . . . . . . . . . . . . . . . . 63
3.2.1 Model Constructions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
3.2.2 Problem Formulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
3.3 Joint Localization and Tracking Design . . . . . . . . . . . . . . . . . . . . . . . . 66
3.3.1 Self-localization Algorithm Design . . . . . . . . . . . . . . . . . . . . . 66
ix
x Contents
xiii
xiv Symbols and Acronyms
Ocean covers 70.8% of the earth’s surface, and it plays an important role in supporting
system of life on the earth. Nonetheless, more than 80% of the ocean’s volume
remains unmapped, unobserved, and unexplored. In order to understand and explore
the ocean, strong ocean observation ability is required to offer technical support. For
that reason, many remote-sensing instruments have been deployed to monitor the
ocean; however, they are effective only for surface monitoring. As for this issue, some
underwater sensor instruments, including sonar array, multibeam swath bathymeter,
and acoustic doppler current profiler, have been deployed to form a three-dimensional
underwater sensor network (USN) for online environment observation. Compared
with the remote-sensing technology, USN has the advantages of increased space-time
cover ability, reduced failure rate, and improved real-time characteristic. However,
USN lacks the necessary flexibility and adaptability, which makes it easy to form a
monitoring blind area and limits the improvement of observation ability. From the
functional viewpoint, the underwater observation network should not only have the
online monitoring ability but also make feedback decision based on the environment
variation.
With regard to this, autonomous underwater vehicles (AUVs) with integrated
measurement sensors are emerged as a promising choice for underwater observation
network. Compared with USNs, AUVs have the strong mobility, enhanced intel-
ligence, and high controllability, which enable them to make necessary feedback
decision based on the environment variation. As reported in [1], the global market
for AUVs is estimated to grow from 638 million dollars in 2020 to 1638 million dol-
lars by 2025, at a compound annual growth rate of 20.8%. For more clear description,
Fig. 1.1 presents the architecture of underwater observation network via AUVs. In
such network, underwater sensor instruments in sensing layer, including hydrophone,
sonar, multibeam sounder, sound velocity profiler, current meter, and depth sensor
© The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2021 1
J. Yan et al., Autonomous Underwater Vehicles, Cognitive Intelligence and Robotics,
https://doi.org/10.1007/978-981-16-6096-2_1
2 1 Introduction
Based on this, three sensors are used to assist the localization of AUV, as shown
T
in Fig. 1.2. Specifically, the position of sensor i is denoted as xi = xi , yi , zi for
i ∈ {1, 2, 3}, and the bearing vector is denoted as β = [β1 , β2 , β3 ]T . Thereby, the
AOA measurement can be formulated as
β = θ (x) + ε, (1.1)
where θ (x) = [θ1 (x) , θ2 (x) , θ3 (x)]T is the bearing vector of AUV, ε = [ε1 , ε2 , ε3 ]T
is the measurement noise with zero mean and covariance matrices S = diag
{σ12 , σ22 , σ32 }. Of note, θi (x) = arctan y−yi
x−xi
for i ∈ {1, 2, 3}.
With (1.1), the maximum likelihood (ML) estimation of AUV is built, i.e.,
1
(x̂, ŷ) = arg min [β − θ (x)]T S−1 [β − θ (x)]
2
1 (βi − θi (x))2
3
= arg min . (1.2)
2 i=1 σi2
Thus, the position of AUV can be obtained by least square method, i.e.,
where A = [sinβ1 , − cos β1 ; sinβ2 , − cos β2 ; sinβ3 , − cos β3 ], R = diag{r12 , r22 , r32 }
and b = [x1 sinβ1 − y1 cos β1 , x2 sinβ2 − y2 cos β2 , x3 sinβ3 − y3 cos β3 ]T .
The TOA-based localization can be divided into one-way propagation time measure-
ment and roundtrip propagation time measurement, as depicted in Fig. 1.3. Specifi-
cally, the main idea of TOA measurement is to use the propagation time measurement
1.2 An Overview of the Existing Works 5
to acquire the relative distance between AUV and sensors. Inspired by this, an AUV-
assisted joint localization and time synchronization scheme was designed in [4]. In
[5], a TOA-based synchronization method was employed to improve localization
accuracy. Also of relevance, the other TOA-based localization scheme was proposed
in [6]. In the following, a brief introduction on TOA-based localization is presented.
As described in Fig. 1.3a, the TOA between AUV and sensor i under one-way
propagation time measurement can be constructed as
Ti = ti − ta ,
1
= xi − x , i ∈ {1, 2, 3, 4}, (1.6)
c
where ta denotes the time when AUV sends location request, ti denotes the time when
sensor i ∈ {1, 2, 3, 4} receives the message from AUV, c is the propagation speed of
acoustic wave, xi is the position vector of sensor i ∈ {1, 2, 3, 4}, and x is the position
vector of AUV.
Similarly, the TOA between AUV and sensor i under roundtrip propagation time
measurement can be constructed according to Fig. 1.3b, i.e.,
and w = [ε1 , ε2 , ε3 , ε4 ]T , where w is measurement noise with zero mean and covari-
ance matrices R. Thus, the TOA measurement with noise measurement can be rewrit-
ten as T̃ = h(x) + w. In this context, the localization optimization problem of AUV
can be given as
−1
x̂ = arg min{[T̃ − h(x)]T R [T̃ − h(x)]}. (1.8)
As shown in Fig. 1.4, the time measurement difference can be acquired through
propagation time measurement. Clearly, the position of AUV can be located with
the help of four sensor nodes. Based on this idea, some TDOA-based localization
schemes can be found in [7, 8]. In the following, a brief introduction on TDOA-based
localization is given.
According to Fig. 1.4, for i ∈ {1, 2, 3, 4} and j = i, the TDOA measurement
between sensor i and j can be constructed as
ti,j = ti − tj ,
1
= xi − x − xj − x , (1.9)
c
where ti is the time when sensor i receives the message from AUV, xi is the position
vector of sensor i ∈ {1, 2, 3, 4}, and x is the position vector of AUV.
Considering the existence of measurement noise, the noisy TDOA measurement
of ti,j can be formulated as
where ni,j denotes the measurement noise, and it is an independent Gaussian random
variable with zero mean.
In view of (1.9) and (1.10), TDOA measurements can be stacked into
⎡ ⎤ ⎡ ⎤ ⎡ ⎤
ct̃1,2 x1 − x − x2 − x cε1,2
⎢ ct̃1,3 ⎥ ⎢ x1 − x − x3 − x ⎥ ⎢ cε1,3 ⎥
⎢ ⎥ ⎢ ⎥ ⎢ ⎥
⎢ ct̃1,4 ⎥ ⎢ x1 − x − x4 − x ⎥ ⎢ cε1,4 ⎥
⎢ ⎥=⎢ ⎥+⎢ ⎥
⎢ ct̃2,3 ⎥ ⎢ x2 − x − x3 − x ⎥ ⎢ cε2,3 ⎥
⎢ ⎥ ⎢ ⎥ ⎢ ⎥ (1.11)
⎣ ct̃2,4 ⎦ ⎣ x2 − x − x4 − x ⎦ ⎣ cε2,4 ⎦
ct̃3,4 x3 − x − x4 − x cε3,4
:= t̃ := ϕ(x) := w̄
where t̃ is TDOA measurement vector, ϕ(x) is nonlinear vector function about
AUV, and w̄ is noise with zero mean and covariance matrix S.
With (1.11), the localization optimization problem is constructed as
−1
x̂ = arg min{[t̃ − ϕ(x)]T S [t̃ − ϕ(x)]}. (1.12)
It is clear that there is no closed solution to (1.12) since the nonlinear function of
ϕ(x). By using Taylor series, one can linearize ϕ(x) around a reference point x0 , i.e.,
where ∇ϕ(x0 ) denotes the partial derivative of ϕ(x) at the value x0 . Referring to [9],
a recursive solution of (1.12) can be acquired, i.e.,
−1 −1
x̂k+1 = x̂k + (∇ T ϕ(xk )S ∇ϕ(xk ))−1 ∇ T ϕ(xk )S [t̃ − ϕ(xk )]. (1.14)
RSS-based localization is to use the RSS to obtain the relative distance between
AUV and sensors, as shown in Fig. 1.5. With the knowledge of RSS measurement,
some RSS-based localization methods were developed in [10, 11]. Next, a brief
introduction on RSS-based localization is presented.
The RSS in dBm on the AUV can be formulated as
di
Pi (di ) = P0 (d0 ) − 10η log10 − γ (di − d0 ) + ni , (1.15)
d0
where Pi (d ) denotes the received power of sensor i for i ∈ {1, 2, 3, 4}, P0 (d0 ) denotes
the known reference power value at a reference distance d0 from AUV, di denotes
8 1 Introduction
the distance between AUV and sensor i, η denotes the path loss exponent, and γ
denotes the frequency-dependent medium absorption. ni is the zero mean Gaussian
measurement noise with variance σ 2 .
It is noted that γ in (1.15) can be given as
f2 f2
γ = 0.11 + 44 + 2.75 × 10−4 f 2 + 0.003, (1.16)
1+f 2 4100 + f 2
∂ ln p(P;x)
where ∂x
denotes the gradient of the log-likelihood function with respect to x,
∂ ln p(P;x)
2
and ∂x∂xT
denotes the second derivation with respect to x.
Typically, the persistent tracking controllers for a single AUV can be mainly clas-
sified into the following five types: (1) proportion integration differentiation (PID)
controller; (2) fuzzy controller; (3) adaptive controller; (4) sliding mode controller;
and (5) neural network controller.
PID control has attracted extensive attentions due to its simplicity, easy implementa-
tion and superior performance. The linear PID controller is composed of proportional
element P, integral element I, and differential element D. Let η denote the position
vector with coordinate in the inertial reference frame and ηd denote the desired tra-
jectory. Define the tracking error as
The PID controller calculates the tracking error e(t), generates a correction based
on the proportional, integral, and differential elements, and attempts to minimize the
error over time. The PID control scheme is a linear combination of each element,
which can be expressed as
t
de(t)
τ (t) = kP e(t) + kI e(ς )dς + kD , (1.21)
dt
0
where τ is the torque force acting on the vehicle. kP > 0, kI > 0 and kD > 0 denote
the proportional, integral and differential parameters, respectively. The control struc-
ture is shown in Fig. 1.6. The P element, I element, and D element correspond to
current error, historic cumulative error, and the future trend of error, respectively. The
control coefficients kP , kI and kD have great influence on the system performance.
However, in the conventional PID controllers, the adjustment of these parameters
depends on personal experience. Moreover, linear fixed-gain PID controller may not
provide satisfactory performance. In order to solve the issue, a sector-bounded gain
function and two nonlinear tracking differentiators were added to the control scheme
[12]. In [13], the control analysis of nonlinear PID with time-varying parameters
was extended to state feedback, tracking, and integral control. However, the results
in [12, 13] require the exact knowledge about the process to achieve better behav-
10 1 Introduction
ior. Although some excellent optimization methods can be used to optimize the PID
parameters, they tend to converge to the local optimal solution, which leads to opti-
mization failure or poor results. In [14], in order to achieve navigation control of
AUV, the PID coefficients were tuned by particle swarm optimization (PSO) algo-
rithm, which can effectively find the global optimal solution. A fuzzy self-adapting
PID controller was presented in [15]. Considering trajectory tracking control prob-
lem of ocean current disturbed AUV, a combination of the fuzzy logic control (FLC)
and PID was proposed in [16]. The PID parameters are processed under the supervi-
sion of the FLC system. Reference [17] proposed a new control structure, including a
feedforward part to compensate external disturbance and unmodeled dynamics, and
a PID-like fuzzy logic feedback part to improve the tracking performance. Consider-
ing the input constrained problem of AUV, Ref. [18] incorporated a model reference
adaptive PID control with anti-windup compensator. A whole range of nonlinear
functions to replace the fixed gain in traditional PID controller were introduced in
[19].
Classical control theory usually requires a mathematical AUV model for design-
ing the controller. The inaccuracy of mathematical modeling of the AUV usually
degrades the performance of the controller. Fuzzy logic controllers (FLCs) offer a
key advantage over traditional control systems. That is, they do not require mathe-
matical models of the AUV. FLCs have the programming capability of human control
behavior and possess the robustness to uncertainties. It can incorporate human-like
thinking into the automatic control system. Fuzzy logic models are universal approx-
imators and can be built based on input–output data samples, expert experience, or
both. A fuzzy logic system includes: the defuzzifier, the knowledge base, the fuzzy
inference engine working on fuzzy rules, and the fuzzifier. The knowledge base for
the FLS includes a series of fuzzy if–then rules. The block diagram is shown in
1.2 An Overview of the Existing Works 11
Fig. 1.7. The primary steps taken by the fuzzy controller can be detailed as: (i) Select
the input and output variables of the fuzzy controller, such that the range conversion
of the variables is performed, (ii) decide the fuzzy language values of each variable,
and then the corresponding membership function is defined, (iii) establish fuzzy con-
trol rules or algorithms, and (iv) determine the fuzzy reasoning and defuzzification
methods. The fuzzy system can be adopted to approximate the system uncertainties.
Let Fil be the fuzzy sets for a state xi , and define its membership function as μFil (xi )
with i = 1, . . . , n. The knowledge base consists of the if–then inference rules.
Rule l: IF x1 is F1l and x2 is F2l and … and xn is Fnl
THEN y is G l , l = 1, 2, . . . , N .
Based on this, the fuzzy system is denoted as
N n
l=1 ȳl i=1 μFil (xi )
y(x) = , (1.22)
N n
l=1 i=1 μFi
l (xi )
where ȳl = maxy∈R μG l (y) with μG l (y) being membership functions, and x =
[x1 , . . . , xn ]T . Let
T
= [1 , . . . , N ] = [y1 , . . . , yN ], and β(x) = [β1 (x), . . . ,
n
μF l (xi )
satisfying β(x) ≤ β ∗ with a positive constant
i=1
βN (x)]T with βl = N n i ,
l=1 (i=1 μF l (xi ))
i
β ∗ . The fuzzy system is expressed as y(x) = T β(x).
Quite recently, considerable attention has been paid to fuzzy tracking control
of AUV systems. In [20], a simplified and nonlinear single-input fuzzy controller
was developed to track the guidance speeds. In [21], a robust disturbance observer-
based fuzzy control scheme was proposed. The methods utilize the benefits of fuzzy
proportional-integral uncertainty estimator, nonlinear disturbance observer, and feed
forward control term. Reference [22] reconstructed the local desired trajectory based
on a special single-input fuzzy model. A good tracking performance can be achieved
even if thruster saturation occurs. Reference [23] presented a robust AUV docking
approach, which incorporated a Tagaki–Sugeno–Kang (TSK) fuzzy inference sys-
tem. The TSK system can assist the vehicle with high level guidance maneuvers in
the form of fuzzified commanded heading and speed vector fields. In [24], an adap-
12 1 Introduction
Lots of control systems are based on a good understanding of the dynamics of under-
water vehicle and its environment. However, the system parameters are difficult to
measure and the basic physical processes are not fully understood. For a cable-driven
underwater vehicle, periodic waves accompanied with strong currents may excite the
motion of the vehicle during teleoperation in the same way as in a mass-spring sys-
tem, but with nonlinear particularities due to the hydrodynamic drag force on the
cable and its tension-shape characteristic. The cable-driven structure brings a lot
of nonlinearities and disturbance to the system, making its dynamic behavior quite
complex. In addition, many operators of low cost underwater vehicle have little or no
access to hydrodynamic testing facilities, thus leading to very little knowledge of the
vehicle’s model parameters. It is widely acknowledged that the accuracy of conven-
tional tracking control approaches in high-speed operations is largely affected by the
parameter uncertainties. The above sensitivity is especially severe for direct-drive
underwater vehicles, for which no gear reduction is available. In order to mitigate the
adverse effects of parametric uncertainties in the dynamics of an underwater vehicle
system, adaptive control methods are being actively researched. Thereby, the devel-
opment of effective adaptive controllers reflects an important step toward versatile
applications of high-speed and high-precision vehicles, including the AUVs.
For several years, great effort has been devoted to the study of adaptive control
for underwater vehicles. The control diagram is shown in Fig. 1.8. To be specific,
Ref. [25] presented a 6-DOF controller for AUV, which is based on PD scheme with
an adaptive compensation of the dynamics. In order to control the underwater vehicle
without requiring a priory knowledge of it, the control law was constructed in [26]
based on the adaptive speed gradient. Besides that, the performance of underwater
vehicles with command governor adaptive control was experimentally validated in
[27].
Update Law
1.2 An Overview of the Existing Works 13
Neural networks (NNs) can be regarded as a block-box estimator for a general class of
systems. One of the greatest advantages of using a neural network-based controller
is its self-learning ability where one do not need a large prior information of the
system to be controlled. Multilayer neural networks can approximate any continuous
functions as accurate as possible [28].
NNs can approximate continuous nonlinear maps. For a suitable approximation of
unknown nonlinear functions, several NN architectures are currently available. With
the universal approximation property of NNs, a continuous function f (Z) ∈ C(S)
within a compact subset S can be approximated as
where Z ∈ Z is the input vector, W is the target weight matrix, φ(Z) denotes the
vector of activation functions, ε(Z) is the NN functional approximation error vector.
The actual NN output is defined as
Hidden
Input
Output
ηd η
AUV model
NN Controller
Reaching Phase
Sliding
Manifold
O
Sliding Phase
Generally, sliding mode control makes the motions to be composed of two phases:
the hitting (or reaching) phase and the sliding phase, as shown in Fig. 1.10. In the
hitting phase, the trajectory of the system starting from everywhere is driven to a
sliding manifold. In the sliding phase, the trajectory is driven to the equilibrium point
along the sliding manifold. The behavior of the system is determined by the sliding
manifold during the sliding mode. The sliding mode control design is a procedure
with two stages: (i) choosing an appropriate manifold as the sliding surface, such
that the reduced dynamics have good performance and convergence properties and
are independent of the perturbations; (ii) designing a discontinuous controller with
the task of driving the trajectories of the system to the prescribed sliding surface
in finite time, and keeping them on it. In spite of claimed interesting properties,
the sliding mode control has a specific disadvantage. The discontinuous nature of
the control law creates “chattering” which may excite unmodeled high-frequency
vibrations of the controlled system, which degrades the performance and may lead
to instability. The first generation of sliding surface is linear hyperplane, which only
guarantees the asymptotical convergence of the system states to the equilibrium
point. In order to achieve finite-time convergence, the terminal sliding mode control
(TSMC) has been widely adopted. In spite of several advantages (finite-time control
performance and high precision), the TSCM approach suffers from slow convergence
speed and intrinsic singular problem due to using fractional power functions. Later,
the evolution of the sliding mode control theory mainly focuses on eliminating the
chattering, singularity, and achieving fast convergence.
Many researchers have investigated the tracking control problem for underwater
vehicle systems using sliding mode technique. Qiao and Zhang [33] proposed an
adaptive second-order fast non-singular terminal sliding mode controller for AUVs
with dynamic uncertainties and environmental disturbances. Based on integral ter-
minal sliding mode (ITSM) and fast ITSM, [34] developed two novel double-looped
adaptive ITSM control scheme for unmanned underwater vehicles. The position
tracking error converges to zero in finite time.
For some special scenarios, a single AUV cannot accomplish the complex and
dynamic missions, e.g., the pursuit of a mobile target and the transport of a payload. In
this regard, the formation control of multiple AUVs is emerged as a promising solu-
tion, where AUVs cooperatively maintain a desired spatial pattern during the whole
tracking procedure. Typically, the architecture of AUV formation can be divided into
the following three types: centralized way, distributed way, and hierarchical way. In
the following, the related works on the above three types are presented.
16 1 Introduction
In distributed way, AUVs can exchange information of environment and self states,
as depicted by Fig. 1.12. In order to enable AUVs keep a desired chape by distributed
way, each AUV requires to share its local information with its neighboring AUVs.
With the collected neighboring information, each AUV decides its formation con-
troller, such that each AUV can maintain a constant distance and the same speed with
its neighbor AUVs.
In view of this, some distributed formation controllers have been developed for
AUVs. For instance, a feedback linearization-based distributed formation controller
was developed in [38]. In [39], a disturbance observer was employed into the dis-
AUV
State AUV 1 AUV n
1.2 An Overview of the Existing Works 17
Communication Channel
Controller 1 AUV 1
Target
State Controller 1 AUV 2
AUV
State
Controller n AUV n
tributed formation control of AUVs, which can guarantee fixed-time stability of the
formation shape. Besides that, the neural adaptive technology was adopted in [40]
to enable leader-following distributed formation. More recently, the learning-based
technologies have been applied to the distributed formation of AUVs. For example,
a formation learning scheme was designed in [41], in which the radial basis function
neural networks were conducted to identify model uncertainty. In [42], a model-free
formation controller was designed in [42] to achieve distributed containment for-
mation of AUVs. Compared with the centralized formation solution, the distributed
formation controllers have the advantages of better robustness, improved scalability,
and reduced computing-communicating complexity.
Target State
Centralized Formation Controller
AUV
AUV
State AUV 1
n-m
Communication Channel
Controller Controller
n-m+1 n
AUV AUV
AUV n-m+1 n
State
In underwater environment, the radio and optical waves are strongly absorbed, which
can only travel to short distances as illustrated by Table 1.1. Alternatively, the acoustic
waves can transmit a signal over distances greater than 100 m, and hence, acous-
tic communication emerges as the most efficient communication way for AUVs
[48]. Compared with the terrestrial vehicles, the following weak communication of
underwater acoustic communication and the model uncertainty of AUVs make it
challenging to achieve the localization, tracking, and formation tasks for AUVs.
Introduction (Chapter 1)
Joint design
Rigid graph Slide mode MPCM Finite time Tracking of Tracking and
based based joint based joint tracking of AUV with formation for
asynchronous localization localization AUV with time delay multiple
localization of and tracking and tracking model and input AUVs with
AUVs of AUV of AUV uncertainty saturation time delay
(Chapter 2) (Chapter 3) (Chapter 4) (Chapters 5-6) (Chapter 7) (Chapter 8)
localization approach for AUVs. The localization approach can be divided into two
phases, i.e., motion prediction and persistent localization. An unscented transform-
based localization estimator is developed in the first phase to acquire the initial
position of AUV, through which rigid graph-based localization scheme is designed
in the second phase to enable ubiquitously localization. For the proposed solution,
the duty-cycle mechanism and unscented Kalman filtering are jointly employed to
improve the localization accuracy and prolong the network lifetime.
Chapters 3 and 4 study the joint localization and tracking issue for a single AUV.
To be specific, Chap. 3 develops a joint localization and tracking solution for an AUV,
subject to asynchronous clocks in physical channel and state disturbances in control
channel. An asynchronous localization algorithm is designed to estimate the posi-
tion of an AUV, through which a model-free tracking controller is provided to drive
AUV to the reference trajectory. Of note, the analytical solutions, i.e., time-delay
estimation and dynamic gain strategies, are jointly employed to reduce the influence
of state disturbances. However, due to the large scale and complex characteristic
of large-scale infrastructure systems, the analytical solution to system dynamics is
generally unavailable. In this context, Chap. 4 develops a reinforcement learning
(RL)-based joint localization and tracking solution for AUV, where the multivariate
probabilistic collocation method (M-PCM) is adopted to evaluate the model uncer-
tainty. Compared with the analytical solutions, the tracking solution in Chap. 4 can
accurately estimate the out mean and effectively achieve the tracking task.
With the assumption that the position information of AUVs can be accurately
acquired, Chap. 5 develops a finite-time tracking controller for a single AUV. Par-
ticularly, a robust differentiator is designed to estimate the uncertainty parameters
1.3 Research Challenges and Contributions of This Book 21
of AUV model. Then, a finite-time tracking control law is designed by using the
non-singular fast terminal sliding mode control technique. Followed by this, Chap. 6
considers the unknown velocity measurements, internal model uncertainties, and
external disturbances. Then, a finite-time tracking solution without velocity mea-
surement is developed for AUV, where position estimator and velocity observer are
both provided.
Chapter 7 further studies the communication delay and actuator saturation con-
straints, such that a model-free proportional-derivative tracking controller is designed
to enforce the position tracking. By using Lyapunov–Krasovskii functions and linear
matrix inequalities, the stability conditions and domain of attraction (DOA) are both
derived. The designed tracking controller in this chapter can guarantee asymptotic
stability, and the DOA can be significantly enlarged as compared with the other
exiting ones.
Chapter 8 considers the communication delays of AUVs. Then, a single-AUV
tracking controller is first designed to state synchronization of the tracking control.
Based on this, a distributed formation controller is developed to enable effective
formation of multiple AUVs. For the developed tracking and formation controllers,
Lyapunov–Krasovskii functions are conducted to analyze the stability. Based on this,
the sufficient conditions for asymptotic stability with time-varying communication
delays are presented.
Chapter 9 provides the future research direction on the localization, tracking,
and formation for AUVs, including the network architecture, channel prediction for
communication support, and optimization control.
References
10. S. Chang, Y. Li, Y. He, H. Wang, Target localization in underwater acoustic sensor networks
using RSS measurements. Appl. Sci. 8(2), 1–14 (2018)
11. T. Xu, Y. Hu, B. Zhang, G. Leus, RSS-based sensor localization in underwater acoustic sensor
networks, in Proceedings of 2016 ICASSP, Shanghai, Mar 2016, pp. 3906–3910
12. Y. Su, D. Sun, B. Duan, Design of an enhanced nonlinear PID controller. Mechatronics 15(8),
1005–1024 (2005)
13. B. Armstrong, D. Neevel, T. Kusik, New results in NPID control: tracking, integral control,
friction compensation and experimental results. IEEE Trans. Control Syst. Technol. 9(2), 399–
406 (2001)
14. L. Zhang, L. Zhang, S. Liu, J. Zhou, C. Papavassiliou, Low-level control technology of micro
autonomous underwater vehicle based on intelligent computing. Clust. Comput. 22(4), 8569–
8580 (2018)
15. M. Khodayari, S. Balochian, Modeling and control of autonomous underwater vehicle (AUV)
in heading and depth attitude via self-adaptive fuzzy PID controller. J. Mar. Sci. Technol. 20(3),
559–578 (2015)
16. M. Hammad, A. Elshenawy, M. Singaby, Trajectory following and stabilization control of
fully actuated AUV using inverse kinematics and self-tuning fuzzy PID. PLoS One 12(7),
1–35 (2017)
17. P. Londhe, S. Mohan, B. Patre, L. Waghmare, Robust task-space control of an autonomous
underwater vehicle-manipulator system by PID-like fuzzy control scheme with disturbance
estimator. Ocean Eng. 139(1), 1–13 (2017)
18. P. Sarhadi, A. Noei, A. Khosravi, Model reference adaptive PID control with anti windup
compensator for an autonomous underwater vehicle. Ocean Eng. 83(1), 87–93 (2016)
19. J. Guerrero, J. Torres, V. Creuze, A. Chemori, E. Campos, Saturation based nonlinear PID
control for underwater vehicles: design stability analysis and experiments. Mechatronics 61(1),
96–105 (2019)
20. C. Yu, X. Xiang, L. Lapierre, Q. Zhang, Nonlinear guidance and fuzzy control for three-
dimensional path following of an underactuated autonomous underwater vehicle. Ocean Eng.
146(1), 457–467 (2017)
21. G. Lakhekar, L. Waghmare, R. Roy, Disturbance observer-based fuzzy adapted S-surface con-
troller for spatial trajectory tracking of autonomous underwater vehicle. IEEE Trans. Intell.
Veh. 4(4), 622–636 (2019)
22. X. Liu, M. Zhang, E. Rogers, Trajectory tracking control for autonomous underwater vehicles
based on fuzzy replanning of a local desired trajectory. IEEE Trans. Veh. Technol. 68(12),
11657–11667 (2019)
23. K. Teo, E. An, P. Beaujean, A robust fuzzy autonomous underwater vehicle (AUV) docking
approach for unknown current disturbances. IEEE J. Oceanic Eng. 37(2), 143–155 (2012)
24. Z. Zhang, Y. Wu, Adaptive fuzzy tracking control of autonomous underwater vehicles with
output constraints. IEEE Trans. Fuzzy Syst. 29(5), 1311–1319 (2021)
25. G. Antonelli, S. Chiaverini, N. Sarkar, M. West, Adaptive control of an autonomous underwater
vehicle: experimental results on ODIN. IEEE Trans. Control Syst. Technol. 9(5), 756–765
(2001)
26. M. Jordan, J. Bustamante, Guidance of underwater vehicles with cable tug perturbations under
fixed and adaptive control systems. IEEE J. Ocean. Eng. 33(4), 579–598 (2008)
27. C. Makavita, S. Jayasinghe, H. Nguyen, D. Ranmuthugala, Experimental study of command
governor adaptive control for unmanned underwater vehicles. IEEE Trans. Control Syst. Tech-
nol. 27(1), 332–345 (2019)
28. K. Hornik, M. Stinchcombe, H. White, Multilayer feedforward networks are universal approx-
imators. Neural Netw. 2(5), 359–366 (1989)
29. B. Park, J. Kwon, H. Kim, Neural network-based output feedback control for reference tracking
of underactuated surface vessels. Automatica 77(1), 353–359 (2017)
30. S. Jagannathan, G. Galan, One-layer neural-network controller with preprocessed inputs for
autonomous underwater vehicles. IEEE Trans. Veh. Technol. 52(5), 1342–1355 (2003)
References 23
31. J. Wang, C. Wang, Y. Wei, C. Zhang, Neuroadaptive sliding mode formation control of
autonomous underwater vehicles with uncertain dynamics. IEEE Syst. J. 14(3), 3325–3333
(2020)
32. M. Chen, D. Zhu, Optimal time-consuming path planning for autonomous underwater vehicles
based on a dynamic neural network model in ocean current environments. IEEE Trans. Veh.
Technol. 69(12), 14401–14412 (2020)
33. L. Qiao, W. Zhang, Adaptive second-order fast nonsingular terminal sliding mode tracking
control for fully actuated autonomous underwater vehicles. IEEE J. Ocean. Eng. 44(2), 363–
385 (2019)
34. L. Qiao, W. Zhang, Double-loop integral terminal sliding mode tracking control for UUVs with
adaptive dynamic compensation of uncertainties and disturbances. IEEE J. Ocean. Eng. 44(1),
29–53 (2019)
35. N. Burlutskiy, Y. Touahmi, B. Lee, Power efficient formation configuration for centralized
leader–follower AUVs control. J. Mar. Sci. Technol. 17(3), 315–329 (2012)
36. S. Choi, J. Kim, Three dimensional formation control to pursue an underwater evader utilizing
underwater robots measuring the sound generated from the evader. IEEE Access 7(1), 150720–
150728 (2019)
37. D. Moreno-Salinas, N. Crasta, A. Pascoal, J. Aranda, Formation control of surface marine
vehicles for underwater target tracking using range information, in Proceedings of 2018 CON-
TROLO, Ponta Delgada, June 2018, pp. 201–206
38. H. Liu, Y. Wang, F. Lewis, Robust distributed formation controller design for a group of
unmanned underwater vehicles. IEEE Trans. Syst. Man Cybern. Syst. 51(2), 1215–1223 (2021)
39. Z. Gao, G. Guo, Fixed-time sliding mode formation control of AUVs based on a disturbance
observer. IEEE/CAA J. Autom. Sin. 7(2), 539–545 (2020)
40. J. Wang, C. Wang, Y. Wei, C. Zhang, Filter-backstepping based neural adaptive formation
control of leader-following multiple AUVs in three dimensional space. Ocean Eng. 201(1),
107150–107160 (2020)
41. C. Yuan, S. Licht, H. He, Formation learning control of multiple autonomous underwater
vehicles with heterogeneous nonlinear uncertain dynamics. IEEE Trans. Cybern. 48(10), 2920–
2934 (2018)
42. N. Gu, D. Wang, Z. Peng, T. Li, S. Tong, Model-free containment control of underactuated
surface vessels under switching topologies based on guiding vector fields and data-driven neural
predictors. IEEE Trans. Cybern. (2021). https://doi.org/10.1109/TCYB.2021.3061588
43. Z. Zhou, J. Yuan, W. Zhang, H. Zhao, J. Zhao, Formation control based on a virtual-leader-
follower hierarchical structure for autonomous underwater vehicles. Int. J. Adv. Comput. Tech-
nol. 4(2), 111–121 (2012)
44. D. Wang, J. Yuan, J. Xu, Z. Zhou, Finite-time formation control for autonomous underwater
vehicles based on hierarchical leader-follower. Appl. Mech. Mater. 541–542(1), 1164–1167
(2014)
45. I. Bychkov, A. Davydov, M. Kenzin, N. Maksimkin, N. Nagul, S. Ulyanov, Intelligent control
of autonomous underwater vehicles groups, in Proceedings of 2017 ICCAR, Nagoya, Apr 2017,
pp. 180–183
46. B. Das, B. Subudhi, B. Pati, Co-operative control of a team of autonomous underwater vehicles
in an obstacle-rich environment. J. Mar. Eng. Technol. 15(3), 135–151 (2016)
47. S. Tangirala, R. Kumar, S. Bhattacharyya, M. O’Connor, L. Holloway, Hybrid-model based
hierarchical mission control architecture for autonomous underwater vehicles, in Proceedings
of 2005 ACC, Portland, Jun 2005, pp. 668–673
48. M. Stojanovic, J. Preisig, Underwater acoustic communication channels: propagation models
and statistical characterization. IEEE Commun. Mag. 47(1), 84–89 (2009)
Chapter 2
Rigid Graph-Based Asynchronous
Localization of AUVs
2.1 Introduction
© The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2021 25
J. Yan et al., Autonomous Underwater Vehicles, Cognitive Intelligence and Robotics,
https://doi.org/10.1007/978-981-16-6096-2_2
26 2 Rigid Graph-Based Asynchronous Localization of AUVs
cial intelligence has also been employed to the localization and routing of USNs,
such as [11–16]. As we known, the localization accuracy can be reduced while the
inhomogeneity characteristic of underwater medium is ignored. With regard of the
asynchronous clock and the inhomogeneity characteristic, how to design an accurate
position acquisition strategy for the target persistent localization becomes the first
issue to be solved.
With the estimated position information, sensor nodes require to perform ubiqui-
tous localization for AUV. Typically, a feasible way to ensure ubiquitous localization
is to improve network connectivity. Nevertheless, a good connectivity can add com-
munication energy consumption and reduce network lifetime. To our knowledge, the
communication energy consumption in water is much higher than the one in terres-
trial environment. Besides, it is difficult to replace the batteries mounted on sensor
nodes. For such a scenario, there is a trade-off between localization accuracy and
energy consumption. To balance the above trade-off, some mobility prediction strate-
gies have been developed in [17–20] to predict the mobility of target, through which
duty-cycle mechanisms are employed to save communication energy consumption.
However, target dynamic in [17–20] is reduced to a second-order differential equa-
tion, and it cannot capture the actual model of AUV, due to which AUV dynamic
usually has nonlinear characteristic with multiple freedoms [21]. On the other hand,
some researchers try to incorporate the graph optimization into the trade-off balance
between connectivity maintenance and energy efficiency, e.g., [22, 23]. Although
the works in [22, 23] demonstrate a good trade-off between connectivity mainte-
nance and energy efficiency, the trade-off between localization accuracy and energy
consumption has not been studied. With consideration of the nonlinear characteristic
of AUV, how to employ the graph optimization technology to balance the trade-off
between localization accuracy and energy consumption for the ubiquitous localiza-
tion is still an open issue to be solved.
To address the above challenges, a rigid graph-based solution is proposed in this
chapter to achieve ubiquitous localization of AUV. The localization process is divided
into two stages, i.e., motion prediction and persistent localization. In the first stage,
sensor nodes employ an unscented transform (UT)-based localization estimator to
obtain the initial position of AUV, through which a terminal sliding mode velocity
observer is designed for sensor nodes to predict the mobility trajectory of AUV. Then,
a minimum rigid graph-based localization strategy is proposed in the second stage to
optimize the network topology. Main contributions of this chapter lie in two aspects:
A network architecture including surface buoys, sensor nodes, and AUV is considered
in this chapter, as described in Fig. 2.1.
• Surface Buoys. Surface buoys obtain their time references and positions via elec-
tromagnetic communication with GPS. The function of buoys is to provide self-
position and clock synchronization services for sensor nodes.
• Sensor Nodes. Sensor nodes can make direct communication with surface buoys.
It is assumed that the positions and time clocks of sensor nodes can be preknown by
utilizing some existing technologies, e.g., [26]. The function of sensor nodes is to
track the real-time position of AUV. To save communication energy consumption,
they work in two modes, i.e., high-power activation mode and low-power sleep
mode.
• AUV. AUV is an untethered mobile vehicle with multiple freedoms, whose real-
time position information requires to be located by contacting with sensor nodes.
It is noted that the time clock between AUV and sensor nodes is not well synchro-
nized. In addition, a cooperation scenario is considered here, i.e., AUV periodically
sends an initiator message to the network.
Specifically, the body-fixed reference frame (BRF) and the inertial reference frame
(IRF) are jointly used to describe AUV dynamic. As mentioned in [27], AUV is
intrinsically stable in roll and pitch. Therefore, the velocity vector of AUV in BRF
can be denoted as v0 = [u0 , v0 , w0 , r0 ]T , where u0 , v0 , and w0 represent the linear
velocities in surge, sway, and heave, respectively. In addition, r0 is the angle velocity
of AUV in yaw. Besides that, the position and heading vector of AUV in IRF can be
denoted as η0 = [x0 , y0 , z0 , ϕ0 ]T , where x0 , y0 , and z0 are the positions in X, Y, and
Z axes, respectively. Meanwhile, ϕ0 represents the angle of AUV in yaw. Based on
this, the motion model of AUV can be expressed as
where J ∈ R4×4 is the rotation matrix, M ∈ R4×4 is the inertia matrix, C ∈ R4×4 is
the Coriolis–centripetal matrix, D ∈ R4×4 denotes the damping matrix, and τ ∈ R4
denotes the torque vector. Readers can refer to [28] for further details about the
modeling of AUV.
Define M = J−T MJ−1 , D = J−T DJ−1 , C = J−T [C − MJ−1 J̇]J−1 and τ̄ = J−T τ .
Then, the motion model of AUV in (2.1) can be rewritten as
where Tauv is local time of AUV. In addition, αauv and βauv are clock skew and offset
with respect to real time t, respectively.
A popular sound model [24, 25, 29] is applied to describe the stratification effect.
The sound velocity at depth z̄ can be given as C(z̄) = āz̄ + b̄, where ā denotes the
steepness of sound speed profile, b̄ denotes the sound velocity at surface, and z̄
denotes the depth. In view of this, the acoustic propagation delay between sender
2.2 Network Model and Problem Formulation 29
node S and receiver node R can be calculated by ray tracking approach, as depicted in
Fig. 2.2. Therefore, the propagation delay between sender S and receiver R is given
as
1 1 + sin θS 1 + sin θR
τS,R = − ln − ln , (2.5)
ā cos θS cos θR
where b denotes the bit length of the data packet, Pr is the power consumption,
Tp is the data transmission time, d denotes the transmission distance between two
nodes, and the practical value of λ is generally set as 1.5. Specifically, g(f ) denotes the
2
44f 2 −4 2
absorption coefficient, i.e., g(f ) = 0.11f
1+f 2
+ 4100+f 2 + 2.75 × 10 f + 0.003 where
f is the signal frequency.
30 2 Rigid Graph-Based Asynchronous Localization of AUVs
Based on the above discussions, the problems studied in this chapter can be presented
as follows.
Problem 2.1 (Observer-Based Motion Prediction) In underwater environment,
clock is difficult to reach accurate synchronization and the data transmission path
follows a bent trajectory. In addition, AUV dynamic has nonlinear property with
multiple freedoms. The aforementioned characteristics make it difficult to predict
the trajectory of AUV. To handle this issue, we try to design an UT-based localization
estimator to obtain AUV’s initial position, through which a terminal sliding mode
velocity observer is built to predict the mobility of AUV. This problem can be reduced
to estimate η0 and v0 with the constraints of (2.3), (2.4), and (2.5).
This section develops a rigid graph-based localization approach for AUV, and its
framework is depicted by Fig. 2.3. Accordingly, the localization approach is divided
into two stages, i.e., (1) the observer-based motion prediction and (2) the minimum
rigid graph-based localization strategy.
where τa,i represents the one-way propagation delay between AUV and sensor node
i ∈ {1, . . . , N1 }. Besides that, a,i ∼ N (0, σmea
2
) represents the time measurement
noise between AUV and sensor node i, while a,j ∼ N (0, σmea 2
) represents the time
32 2 Rigid Graph-Based Asynchronous Localization of AUVs
measurement noise between AUV and sensor node j. It is worth mentioning that,
τa,i can be calculated by (2.5), and i,j = a,i − a,j satisfies the distribution of
i,j ∼ N (0, 2σmea ).
2
p̃0 (0) = argmin[(φ − ξ (p0 (0), pi ))T (φ − ξ (p0 (0), pi ))]. (2.10)
p0 (0)
where H(p̃l0 (0)) denotes the Jacobian matrix at p̃l0 (0) and l denotes the iterative
increment. It is noted that the calculation of Jacobian matrix can introduce large
linearization error. To handle this issue, an UT-based method [31] is adopted to
locate H, which consists of the following phases:
(a) (Sigma point selection). Based on the arbitrary prior mean p̄0 and the covariance
matrix Pp̃l0 (0)p̃l0 (0) , one selects seven sigma points, which can be presented as Xsl
with s ∈ {0, . . . , 6};
(b) (Iteration update). Through weight average, p̄0 and Pp̃l0 (0)p̃l0 (0) is updated with
weight Ws , i.e., p̄0 = 6s=0 Ws Xsl and Pp̃l0 (0)p̃l0 (0) = 6s=0 Ws [Xsl − p̄0 ][Xsl −
p̄0 ]T . In view of this, sigma points can be updated. Instantiating each sigma
point through (2.9), one has φ ls = ξ (Xsl , pi ). Accordingly, the prior measure-
l
ment estimation is given as φ̄ = 6s=0 Ws φ ls . Then, the predicted measurement
covariance matrix Pφφ and cross-covariance matrix Pp̃l0 (0)φ can be calculated as
6
l l
Pφφ = Ws [φ ls − φ̄ ][φ ls − φ̄ ]T + R,
s=0
(2.12)
6
l T
Pp̃l0 (0)φ = Ws [Xsl − p̄0 ][φ ls − φ̄ ] .
s=0
Another random document with
no related content on Scribd:
one very important in the animal economy, and having various relation to
the causes and treatment of disease. Keeping in mind then this reference
to the use of water as an internal remedy, diluents may be viewed under
three conditions of probable usefulness:—First, the mere mechanical
effect of quantity of liquid in diluting and washing away matters,
excrementitious or noxious, from the alimentary canal;—secondly, their
influence in modifying certain morbid conditions of the blood;—and thirdly,
their effect upon various functions of secretion and excretion, and
especially upon those of the kidneys and skin * * * The first is an obvious
benefit in many cases, and not to be disdained from any notion of its
vulgar simplicity. It is certain there are many states of the alimentary canal
in which the free use of water at stated times produces good, which
cannot be attained by other or stronger remedies. I have often known the
action of the bowels to be maintained with regularity for a long period,
simply by a tumbler of water, warm or cold, on an empty stomach, in
cases where medicine had almost lost its effect, or become a source only
of distressing irritation. The advantage of such treatment is still more
strongly attested, where the secretions taking place into the intestines, or
the products formed there during digestion, become vitiated in kind. Here
dilution lessens that irritation to the membranes, which we cannot so
readily obviate by other means, and aids in removing the cause from the
body with less distress than any other remedy. In some cases, where
often and largely used, its effect goes farther in actually altering the state
of the secreting surfaces by direct application to them. I mention these
circumstances upon experience, having often obtained much good from
resorting to them in practice, when stronger medicines and ordinary
methods had proved of little avail. Dilution thus used, for example, so as to
act on the contents of the bowels, is beneficial in many dyspeptic cases,
where it is especially an object to avoid needless irritation to the system.
Half-a-pint or more of water taken when fasting, at the temperature most
agreeable to the patient, will often be found to give singular relief to his
morbid sensations. * * * In reference to the foregoing uses of diluents, it is
to be kept in mind that the lining of the alimentary canal is, to all intents, a
surface, as well as the skin, pretty nearly equal in extent; exercising some
similar functions, with others more appropriate to itself, and capable in
many respects of being acted upon in a similar manner. As respects the
subject before us, it is both expedient and correct in many cases to regard
diluents as acting on this internal surface analogously to liquids on the
skin. And I would apply this remark not only to the mechanical effects of
the remedy, but also to their use as the medium for conveying cold to
internal parts; a point [29]of practice which either the simplicity of the
means, or the false alarms besetting it, have hitherto prevented from being
duly regarded.”
Again he writes:—
And again:—
“As I have been treating of this remedy only in its simplest form, I do not
advert to the use of the different mineral waters farther than to state, that
they confirm these general views, separating, as far as can be done, their
effect as diluents from that of the ingredients they contain. The copious
employment of some of them in Continental practice gives room for
observation, which is wanting under our more limited use. I have often
seen five or six pints taken daily for some weeks together (a great part of it
in the morning while fasting), with singular benefit in many cases to the
general health, and most obviously to the state of the secretions. * * *
These courses, however, were always conjoined with ample exercise and
regular habits of life; doubtless influencing much the action of the waters,
and aiding their salutary effect.”
“It is a vital stimulus, and is more essential to our existence than aliment.”
Are these effects consistent with lowering the tone of the stomach?
are they not, on the contrary, the strongest evidence of the tonic
effects of water?
“But the whole assertion regarding thin blood proceeds on grounds that
betray intense ignorance, both of physiology and of the water cure. It
supposes that the whole water imbibed enters into, and remains in the
circulating blood, quasi water, that no chemical transformation of it takes
[30]place in the body at all: this is ignorance of physiology. And it supposes
that all who are treated by water are told to drink the same, and that a
large quantity, without discrimination of the individual cases of disease
presented: this is ignorance of the water cure. So between the horns of
this compound ignorance, and of wilful misrepresentation, we leave the
declaimers about the ‘thinning of the blood.’ ”
In so flagrant a case of thin blood, why has this principal cause been
omitted? It is further curious that this injurious effect of water was
never invented, much less preached, until Hydropathy was found to
be making inconvenient strides in public favour.
Is the reader aware that eighty per cent. of water enters into the
composition of healthy blood, without making any allowance for the
enormous quantity required for the various secretions?
Granting, however, for the sake of argument, that all, and more than
these objectors urge, were true, we still have a kind of feeling that
water is more congenial to the system than prussic acid, or even
iodine. But we may be wrong.
“Tell Barter that his system has lately become the universal practice in the
Southern States, for cholera; and since its adoption, although it is, of
course, but imperfectly carried out, the mortality is not one-fourth.
“Need more be offered upon the subject; and yet with such facts upon
record, ‘hot punch’ is now given to the poor patients in the cholera
hospitals in Limerick. Those pious and angelic Sisters of Mercy, to whom
you have alluded in the Chronicle, never, in all probability, heard or read of
the treatment of cholera as above narrated; but ever attentive and
observant as they are in the performance of their hallowed vocation, they
have not been unmindful of the good effects of cold water. Nature prompts
the sufferer to call for it, and it should be always supplied. In cholera, pure
water is balsamic.
“As to the operation of cold water on the human system in cholera, or the
action of the system on water, I will not presume to pronounce; but I may
say that it is commonly supposed that when the serum (one of the
important constituents of the blood) is exhausted by discharges, collapse
takes place, and the livid hue of the countenance follows; and everybody
has heard of the experimental operation of transfusion of warm water,
combined with albumen and soda, into the veins, to supply the absence of
serum, in order to give the vital current its natural and healthy flow:
whether cold water, from the oxygen it contains, and the necessary heat it
is therefore calculated to impart, is taken up rapidly by the absorbents to
cherish and feed the blood, and fill the channels of circulation, so as to
remove collapse in cholera, I shall leave physiologists to determine; but it
is indisputable that cholera patients have anxiously asked for, and eagerly
swallowed, copious draughts of cold water, till their thirst was allayed,
genial warmth restored, agony banished, and the vital functions vivified
and invigorated.” * * * * *
“I am acquainted with three persons, who, after they had been laid out for
dead, on being washed, previous to interment, in the open court-yard, with
water, to obtain which the ice had been broken, recovered in
consequence, and lived many years. I received from Erycroon, in Turkey,
a letter from our excellent Consul, Mr. Brant, who states that Dr. Dixon, of
that place, was then curing more patients by friction, with ice or snow, than
any other treatment. The same practice is reported to have been the most
effectual in Russia.”
“I never yet saw a patient that did not cry out for cold water; and the
confirmed dram-drinker can, with difficulty, be persuaded to taste his
[33]favourite beverage; he objects more to brandy or punch than the
temperate do; this I have often remarked. I have seen a patient travel for
miles on an open car, through sleet and rain, without any covering, and
drinking cold water on the way, and remarked that he did better than when
treated with brandy, hot tins, &c. In fact, I often saw such patients beg to
be allowed out again, they used to call loudly for cold water. ‘For the love
and honour of God, sir, get us a drink of cold water,’ was no unfrequent
request amongst them, and that pronounced with an earnestness of
manner most truly impressive; but, alas! in 1832, this appeal was always
refused, though in 1849 a step has been taken in a right direction, and it is
allowed, according to the Sisters of Mercy, ‘in small quantities.’ ”
The truth will ere long be acknowledged, that it is our mode of life
that makes us fit subjects for cholera, and that it is our mode of
treating it alone, which makes the disease so dangerous. The wretch
who is cast uncared for in a ditch, exposed to all the inclemency of
the weather, with water alone to quench his burning thirst, has ten
chances to one in favour of his recovery, compared with the well-
cared patient who is dosed with brandy and the favourite specifics of
the apothecary’s shop. If we look at cholera, and divest our minds of
its accustomed mode of treatment, we will find that every symptom
of the disease points to the presence of some highly irritant poison in
the blood; and in the effort to expel this poison, the serum which
contains it, is drained from the system. What, therefore, can be more
rational than to supply the system with the materials of restoration,
by giving water in large quantities, and to stimulate its chemical
combinations by which the caloric of the system shall be restored, by
the influence of fresh air, water drinking, and cold bathing.
“It would be folly, however, to avoid a treatment because it will not for ever
root up your disease in your own convenient time. Look at the destructive
manner in which colchicum reduces a gouty fit, how it approximates the
attacks, and utterly disorganizes the viscera; and then regard what the
water cure is capable of doing, both against individual attacks, and in
reduction of the diathesis, the vital parts meanwhile improving under its
operation; … if it does not utterly cure the gout, at least it does not shorten
the patient’s life as colchicum does.”
“To the patient, and, indeed, to the physician who knows little of
physiology, all this will appear right: the gout is removed, and that is what
[35]was desired. The physician, however, who is a physiologist, will say,
‘True, that irritation which you call gout, has left the extremities, whither it
had been sent by nature to save her noble internal parts. But look to the
signs exhibited by those parts; are they not those of augmented irritation,
at least of irritation of a degree and kind that did not exist so long as the
limbs were pained and inflamed? The fact is, that your colchicum has set
up in the viscera so intense an irritation as to reconcentrate the mischief
within; and the fit is cured, not by ridding the body of the gouty irritation,
but by driving or drawing it in again,’ (thus baffling nature’s efforts at self
relief). ‘Hence the continuance of the dyspeptic symptoms after the fit;
hence, as you will find, the recurrence of another fit ere long, the intervals
becoming less and less, until gouty pain is incessantly in the limbs, and
gouty irritation always in the viscera.’ ”
[36]
Nor will the above results form at all a subject for wonder, when it is
remembered that every natural disease arises either from impurity in
the blood or maldistribution of it, and that all the processes of the
water cure, from the Turkish bath down to the wet sheet, act
powerfully as depurators of the blood and controllers of its
circulation,—attracting it here, and repelling it there, at will.
We know not whether the public will prefer the impartial testimony of
an intelligent observer like Sir Bulwer Lytton, to that of the Allopathic
physician, naturally wedded to his own system and anxious to
sustain it against all intruders; but we may observe, that we never
yet met a physician opposed to Hydropathy, who did not, on
catechising him, exhibit the most absurd ignorance respecting it.
Their chronic idea is that of a person being left to shiver in wet
sheets; and, as a consequence, their chronic note of warning,
accompanied by an ominous shake of the head, consists in, “Don’t
attempt the water cure, or it will kill you.” 17 If medical men would but
see, before they assert, then much value might be attached to their
opinion; but what value can be attached to their opinion about a
system which they will not take the trouble of examining into? How
many orthodox physicians have ever visited Blarney, or any similar
Hydropathic establishment?—The proportion of such visitors (and no
one can form a fair idea of the system without seeing it at work), to
the whole profession would be more than represented by an
infinitesimal fraction.
[Contents]
THE TURKISH BATH. 1