Ebook Autonomous Underwater Vehicles Localization Tracking and Formation Cognitive Intelligence and Robotics Yan Online PDF All Chapter

Autonomous Underwater Vehicles:
Localization, Tracking, and Formation

(Cognitive Intelligence and Robotics)
Yan
Visit to download the full and correct content document:
https://ebookmeta.com/product/autonomous-underwater-vehicles-localization-tracking
-and-formation-cognitive-intelligence-and-robotics-yan/
More products digital (pdf, epub, mobi) instant
download maybe you interests ...
Autonomous Underwater Vehicles Jing Yan Xian Yang

Haiyan Zhao Xiaoyuan Luo Xinping Guan
https://ebookmeta.com/product/autonomous-underwater-vehicles-
jing-yan-xian-yang-haiyan-zhao-xiaoyuan-luo-xinping-guan/
Good Robot, Bad Robot: Dark And Creepy Sides Of

Robotics, Autonomous Vehicles, And AI 1st Edition Jo
Ann Oravec
https://ebookmeta.com/product/good-robot-bad-robot-dark-and-
creepy-sides-of-robotics-autonomous-vehicles-and-ai-1st-edition-
jo-ann-oravec/
Embedded Robotics: From Mobile Robots to Autonomous

Vehicles with Raspberry Pi and Arduino, 4th Edition
Thomas Bräunl
https://ebookmeta.com/product/embedded-robotics-from-mobile-
robots-to-autonomous-vehicles-with-raspberry-pi-and-arduino-4th-
edition-thomas-braunl/
Cambridge IGCSE and O Level History Workbook 2C - Depth

Study: the United States, 1919-41 2nd Edition Benjamin
Harrison
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/
Received Signal Strength Based Target Localization and

Tracking Using Wireless Sensor Networks Satish R
Jondhale R Maheswar Jaime Lloret
https://ebookmeta.com/product/received-signal-strength-based-
target-localization-and-tracking-using-wireless-sensor-networks-
satish-r-jondhale-r-maheswar-jaime-lloret/
Autonomous Driving and Advanced Driver-Assistance

Systems (ADAS): Applications, Development, Legal
Issues, and Testing (Chapman & Hall/CRC Artificial
Intelligence and Robotics Series) 1st Edition Lentin
Joseph (Editor)
https://ebookmeta.com/product/autonomous-driving-and-advanced-
driver-assistance-systems-adas-applications-development-legal-
issues-and-testing-chapman-hall-crc-artificial-intelligence-and-
robotics-series-1st-edition-lent/
Autonomous and Connected Vehicles Network Architectures

from Legacy Networks to Automotive Ethernet 1st Edition
Dominique Paret
https://ebookmeta.com/product/autonomous-and-connected-vehicles-
network-architectures-from-legacy-networks-to-automotive-
ethernet-1st-edition-dominique-paret/
Autonomous Vehicles Volume 1 Using Machine Intelligence

1st Edition A Mary Sowjanya Syed Imran Patel Varshali
Jaiswal Imran Khan Allam Balaram Romil Rawat
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
Jing Yan · Xian Yang · Haiyan Zhao ·

Xiaoyuan Luo · Xinping Guan
Autonomous
Underwater
Vehicles
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.
More information about this series at http://www.springer.com/series/15488

Jing Yan · Xian Yang · Haiyan Zhao ·
Xiaoyuan Luo · Xinping Guan
Autonomous Underwater
Vehicles
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
ISSN 2520-1956 ISSN 2520-1964 (electronic)

ISBN 978-981-16-6095-5 ISBN 978-981-16-6096-2 (eBook)
https://doi.org/10.1007/978-981-16-6096-2
© 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
The results presented in this book mark a substantial contribution to nonlinear

system theory, robotic control theory, and underwater acoustic communication
system. As such, the book will be of great interest to university researchers, scien-
tists, engineers, and graduate students in control science and engineering who wish
to learn about the core principles, methods, algorithms, and applications of AUVs.
Besides that, the practical localization, tracking, and formation schemes presented
in this book can provide the guidance on understanding and exploring the ocean. In
order to understood easy, the knowledge of nonlinear system theory, robotic control
theory, and underwater acoustic communication system is needed.
The book is organized as follows.
Chapter 1 first draws out the role of AUVs in the underwater observation network.
Then, an overview of the existing works is presented, through which the weak acoustic
communication and model uncertainty are given. Finally, the research challenges and
contributions of this book are summarized.
Chapter 2 considers the ubiquitous localization issue for AUVs in the presence of
asynchronous clock and the inhomogeneity characteristic. To this end, a rigid graph-
based localization scheme is developed to balance the trade-off between localization
accuracy and energy consumption for AUVs.
Chapters 3–4 consider the joint localization and tracking issue for a single AUV.
Particularly, Chap. 3 addresses a joint localization and tracking issue for AUV, subject
to asynchronous clocks in physical channel and state disturbances in control channel.
Note that the analytical-based uncertainty evaluation method is adopted in Chap. 3
to capture the uncertainties and 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 view of this, Chap. 4 incorporates
the multivariate probabilistic collocation method into the reinforcement learning
(RL) framework, such that an RL solution is designed for AUV. Compared with the
analytical solutions, the tracking solution in Chap. 4 can accurately estimate the out
mean and effectively achieve the tracking task.
Chapters 5 and 6 consider the finite-time tracking issue for a single AUV, subject
to unknown parameters and external disturbances. A sliding mode differentiator is
developed in Chap. 5 to estimate the uncertain parameters and external disturbances,
such that a finite-time tracking control law is designed to drive the tracking error
to zero. On the basis of this, Chap. 6 jointly considers the unavailable velocity
signals and uncertain model parameters. Then, a finite-time tracking solution without
velocity measurements is developed for AUV, where a buoy-assisted localization
estimator is provided to estimate its position information, and a fast terminal sliding
mode observer is designed to estimate its velocity information.
Chapter 7 further considers the communication delay in cyber-channel and actu-
ator saturation constraints in physical channel, 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 are both presented.
Chapter 8 considers the tracking of a single AUV and the formation of multiple
AUVs, under the constraint of time-varying communication delay. By adopting the
Preface vii
leader–follower strategy, tracking and formation controllers are developed to stabilize

the time-varying AUV system. Besides that, sufficient conditions for the asymptotic
stability are provided to establish the relationship between tracking (or formation)
controller gains and time delay, such that the allowable upper bound of time delay
is calculated.
Chapter 9 presents the future research direction on the localization, tracking,
and formation for AUVs. We have tried to provide complete instructions and
share insights into the underwater localization, tracking, and formation from the
communication and control perspectives.
This book was supported in part by the National Natural Science Foundation of
China under Grants 61873345, 62033011, and 61973263, and by Outstanding Young
Foundation of Hebei Natural Science under Grants 2020203002 and 2021203056. We
would like to acknowledge Prof. Cailian Chen for her value discussions on preparing
material.
Qinhuangdao, China Jing Yan

Qinhuangdao, China Xian Yang
Qinhuangdao, China Haiyan Zhao
Qinhuangdao, China Xiaoyuan Luo
Shanghai, China Xinping Guan
July 2021
Contents
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
3.3 Joint Localization and Tracking Design . . . . . . . . . . . . . . . . . . . . . . . . 66
3.3.1 Self-localization Algorithm Design . . . . . . . . . . . . . . . . . . . . . 66
ix
x Contents
3.3.2 Model-Free Tracking Controller Design . . . . . . . . . . . . . . . . . 70

3.4 Performance Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
3.4.1 Convergence of the Localization Algorithm . . . . . . . . . . . . . . 73
3.4.2 CRLB of the Localization Algorithm . . . . . . . . . . . . . . . . . . . 74
3.4.3 Boundness of the TDE Error . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
3.4.4 Convergence for the Tracking Controller . . . . . . . . . . . . . . . . 78
3.5.2 Experimental Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
3.6 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
4 Joint Localization and Tracking of AUV Via Multivariate
Probabilistic Collocation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
4.2 System Model and Problem Formulation . . . . . . . . . . . . . . . . . . . . . . . 92
4.2.1 Model Constructions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
4.3 Joint Localization and Tracking Design . . . . . . . . . . . . . . . . . . . . . . . . 95
4.3.1 Self-localization Algorithm Design . . . . . . . . . . . . . . . . . . . . . 95
4.3.2 RL-Based Tracking Controller Design . . . . . . . . . . . . . . . . . . 99
4.4 Simulation Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
4.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
5 Finite-Time Tracking Control of AUV with Model Uncertainty . . . . . 113
5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
5.2 System Formulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
5.3 Control Design and Stability Analysis . . . . . . . . . . . . . . . . . . . . . . . . . 116
5.3.1 Tracking Control Without Disturbance . . . . . . . . . . . . . . . . . . 116
5.3.2 Finite-Time Tracking Control with Disturbances . . . . . . . . . . 120
5.4 Simulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125
5.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131
6 Finite-Time Tracking Control of AUV Without Velocity
Measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
6.2 Problem Formulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134
6.3 Finite-Time Tracking Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137
6.3.1 Finite-Time Velocity Observer . . . . . . . . . . . . . . . . . . . . . . . . . 137
6.3.2 A-NFTSM Tracking Controller . . . . . . . . . . . . . . . . . . . . . . . . 142
6.4 Performance Analyses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
6.4.1 Accuracy of the Buoy-Assisted Localization . . . . . . . . . . . . . 144
6.4.2 Stability Analysis for the Velocity Observer . . . . . . . . . . . . . . 144
6.4.3 Stability Analysis for the Tracking Controller . . . . . . . . . . . . 147
Contents xi
6.5 Simulation and Experimental Studies . . . . . . . . . . . . . . . . . . . . . . . . . . 149

6.6 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162
7 Tracking Control of AUV with Time Delay and Input Saturation . . . 165
7.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
7.2 Model and Problem Formulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166
7.3 Controller Design and Stability Analysis . . . . . . . . . . . . . . . . . . . . . . . 168
7.3.1 Design of the Tracking Controller . . . . . . . . . . . . . . . . . . . . . . 168
7.3.2 Stability Condition and DOA Estimation . . . . . . . . . . . . . . . . 168
7.4 Simulation and Experiment Results . . . . . . . . . . . . . . . . . . . . . . . . . . . 175
7.4.2 Implementation and Experimental Studies . . . . . . . . . . . . . . . 179
7.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183
8 Tracking and Formation for Multiple AUVs with Time Delay . . . . . . . 185
8.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185
8.2 Problem Formulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186
8.3 Controller Design and Stability Analysis . . . . . . . . . . . . . . . . . . . . . . . 188
8.3.1 Tracking Control for Single-AUV System . . . . . . . . . . . . . . . 188
8.3.2 Formation Control for Multi-AUV System . . . . . . . . . . . . . . . 191
8.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 202
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 204
9 Future Research Directions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207
9.1 Space-Air-Ground-Sea Network Architecture . . . . . . . . . . . . . . . . . . . 207
9.2 Channel Prediction for Communication Support . . . . . . . . . . . . . . . . 208
9.3 Model-Free Optimization Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 210
Symbols and Acronyms
R Field of real numbers

Rn n-dimensional real Euclidean space
Rn×m Space of n × m real matrices
I Identity matrix
0 Zero matrix
A System matrix
A−1 Inverse of matrix A
AT Transpose of matrix A
lim Limit
min Minimum
max Maximum
argmin f Value of the variable that minimizes function f
argmax f Value of the variable that maximizes function f
tr (A) Trace of matrix A
rank (A) Rank of matrix A
· Euclidean norm
∇f The gradient of function f
∀ For all
∈
Belong to
Sum
Product operation
E{·} Mathematical expectation operator
var{·} Mathematical variance operator
(A)s The sth column of matrix A
diag{·} Diagonal matrix
→ Tend to, or mapping to (case sensitive)
|x| Absolute value (or modulus) of x
sgn(·) Signum function
sign(x) Sign of x
sat() Saturation function with unit limits
co(·) Convex hull of a set
xiii
xiv Symbols and Acronyms
λ(A) Eigenvalue of matrix A

λmax (A) Maximum eigenvalue of matrix A
0n×m Zero matrix of dimension n × m
3-D Three-dimensional
6DOF Six degrees of freedom
AOA Angle of arrival
AUVs Autonomous underwater vehicles
BRF Body-fixed reference frame
CRLB Cramér-rao lower bound
CSMC Conventional sliding mode controller
CTSMC Conventional terminal sliding mode controller
DOA Domain of attraction
DOF Degrees of freedom
DVL Doppler velocity log
FIM Fisher information matrix
FLC Fuzzy logic control
FLCs Fuzzy logic controllers
GPS Global positioning system
HOTL Human-on-the-loop
IMU Inertial measurement unit
IRF Inertial reference frame
ITSM Integral terminal sliding mode
LBL Long-baseline
LMIs Linear matrix inequalities
LMPC Lyapunov-based model predictive control
ML Maximum likelihood
M-PCM Multivariate probabilistic collocation method
MSE Mean square error
NNs Neural networks
NTSMC Non-singular terminal sliding mode controller
PCRLB Posterior cramér-rao lower bound
PD Proportional-derivative
PFIM Predicted Fisher information matrix
PID Proportion integration differentiation
PSO Particle swarm optimization
RL Reinforcement learning
RSS Received signal strength
SBL Short-baseline
SLMP Scalable localization method with mobility prediction
SMC Sliding mode control
TDE Time-delay estimation
TDOA Time difference of arrival
TOA Time of arrival
TSK Tagaki–Sugeno–Kang
TSMC Terminal sliding mode control
Symbols and Acronyms xv
UAVs Unmanned aerial vehicles

UKF Unscented Kalman filtering
USNs Underwater sensor networks
UT Unscented transform
UWB Ultra-wideband
Chapter 1
Introduction
1.1 Underwater Observation Network
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
Fig. 1.1 Architecture of underwater observation network via AUVs
are cooperatively assisted to support the decision-making process of AUVs. Based

on this, AUVs continuously feed back to the control center through the assistance
of fixed platforms in fusion layer. With the reverse communication links, operator in
control layer makes and transmits its local mission decision to AUVs through optical
and acoustic communication channels.
In the architecture, the localization, tracking, and formation of AUVs are of great
importance to the underwater observation network. To be specific, the ubiquitous
localization of AUVs is the prerequisite of persistent tracking and cooperative forma-
tion. Meanwhile, the persistent tracking and cooperative formation are the technical
cores for a more wide range of application-specific missions of AUVs, especially
when the static sensor nodes cannot perform the dynamic monitoring mission. How-
ever, 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 view of this, we first provide an overview of the existing works on
localization, tracking, and formation of AUVs. Followed by this, the weak commu-
nication of underwater acoustic communication and the model uncertainty of AUVs
are presented. Based on this, the challenging works and contributions of this book
are summarized.
1.2 An Overview of the Existing Works 3
1.2 An Overview of the Existing Works
1.2.1 Ubiquitous Localization of AUVs
According to the different measurement ways, the ubiquitous localization schemes

for AUVs can be classified into the following four categories: (1) angle of arrival
(AOA)-based localization; (2) time of arrival (TOA)-based localization; (3) time
difference of arrival (TDOA)-based localization; (4) received signal strength (RSS)-
based localization.
1.2.1.1 AOA-Based Localization
The positions of AUVs can be estimated by AOA measurement. Particularly, each

AUV obtains the direction of signal arrival at transmitter of sensor node by receiver
antenna, such that the relative azimuth between AUV and sensor nodes can be cal-
culated. Based on this, optimization algorithms are employed to locate the positions
of AUVs. Without loss of generality, it is assumed that the beam of receiver antenna
is rotating, and the direction of the maximum received signal strength is regarded as
the direction of AUV. With AOA measurement, a localization scheme was developed
in [2]. In [3], the accurate distance can be acquired through AoA measurement. In
the following, a brief introduction on AOA-based localization is given.
T
Denote the position of AUV as x = x, y, z . Without loss of generality, we
assume that the depth of AUA can be acquired by depth sensor (i.e., z is preknown).
Fig. 1.2 An illustration of

the AOA measurement
4 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
According to Stanfield’s method, it is assumed that the measurement noise is

small enough; one has εi ≈ sin εi . Based on this, one has
1 sin2 (βi − θi (x))

3
(x̂, ŷ) = arg min . (1.3)
2 i=1 σi2

Define ri = (x − xi )2 + (y − yi )2 . By using trigonometric transformation, Eq.
(1.3) can be rearranged as
1 [(y − yi ) cos βi − (x − xi ) sin βi ]2

3
(x̂, ŷ) = arg min . (1.4)
2 i=1 σi2 ri2
Thus, the position of AUV can be obtained by least square method, i.e.,
(x̂, ŷ) = (AT R−1 S−1 A)−1 AT R−1 S−1 b, (1.5)
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 .
1.2.1.2 TOA-Based Localization
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
Fig. 1.3 An illustration of the TOA measurement
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.,
Ti = (ti,a − ta,a ) − τi ,

2
= xi − x , i ∈ {1, 2, 3, 4}, (1.7)
c
where ta,a is the time when AUV sends location request, ti,a is the time when AUV
receives the reply message from sensor i, and τi is the delay for the sensor i to handle
the message. It is noted that the delay is either known by a priori calibration or
measured to be subtracted.
With (1.6) and (1.7), we define T̃ = [cT1 , cT2 , cT3 , cT4 ]T (or T̃ =
[ 2c T1 , 2c T2 , 2c T3 , 2c T4 ]T ), h(x) = [x1 − x , x2 − x , x3 − x , x4 − x]T ,
6 1 Introduction
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)
Based on the principle of trilateration, the localization optimization problem can

be solved, i.e., the position of AUV can be located.
1.2.1.3 TDOA-Based Localization
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

TDOA measurement
t̃i,j = ti,j + ni,j , (1.10)
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.,
ϕ(x) ≈ ϕ(x0 ) + ∇ϕ(x0 )(x − x0 ), (1.13)
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)
1.2.1.4 RSS-Based Localization
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

RSS measurement
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
where f is the frequency of acoustic communication.

Let P = [P1 (d1 ), P2 (d2 ), P3 (d3 ), P4 (d4 )]T , g(x) = [P0 (d0 ) − 10η log10 ( dd01 ) − γ
(d1 − d0 ), P0 (d0 ) − 10η log10 ( dd20 ) − γ (d2 − d0 ), P0 (d0 ) − 10η log10 ( dd03 ) − γ (d3 −
d0 ), P0 (d0 ) − 10η log10 ( dd04 ) − γ (d4 − d0 )]T , and n = [n1 , n2 , n3 , n4 ]T . Thereby, the
likelihood function can be formulated as

1 1
p(P; x) = √ exp − 2 P − g(x) . 2
(1.17)
2π σ 2σ
Based on this, the localization optimization problem of AUV is given as
x̂ = arg max p(P; x). (1.18)
It is noted that the optimization problem in (1.18) is highly non-convex. Thus, it

is difficult to solve analytically. To handle this optimization problem, the Newton–
Raphson method is introduced, i.e.,
−1
∂ 2 ln p(P; x) ∂ ln p(P; x)
x̂k+1 = x̂k − , (1.19)
∂x∂xT ∂x x=x̂k
∂ 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.
1.2.2 Persistent Tracking Control for a Single AUV
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.
1.2.2.1 PID Tracking 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
e(t) = ηd (t) − η(t). (1.20)
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
Fig. 1.6 PID control for AUV tracking
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].
1.2.2.2 Fuzzy Tracking Controller
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
Fig. 1.7 Fuzzy control for AUV tracking
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
tive fuzzy-based controller was designed to guarantee tracking performance, where

fuzzy system was employed to approximate model uncertainties.
1.2.2.3 Adaptive Tracking Controller
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].
Fig. 1.8 Adaptive control

for AUV tracking
ηd e Adaptive η
Controller AUV model
Update Law
Adaptive control, nevertheless, as a branch of system theory, is curtailed in indus-

trial applications. First, because the parameter adaptation is in a time varying and
nonlinear manner, high-frequency oscillations in the control signal at high adaptive
gain values often occur, while acceptable transient tracking performance can not be
achieved at low adaptive gain values. Second, the adaptive control method is compli-
cated and cannot be used for unified and standardized controllers. Third, it is difficult
to acquire and improve a general solution with the adaptive control theory.
1.2.2.4 Neural Network Tracking Controller
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
f (Z) = WT φ(Z) + ε(Z), (1.23)
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
f̂(Z) = ŴT φ(Z), (1.24)
where Ŵ is the actual weight matrix.

It is should be emphasized that neural network method has been used to investi-
gate the tracking control problems of underwater vehicle systems in many existing
literature. The control structure is shown in Fig. 1.9. Reference [29] proposed a neural
network-based output feedback control algorithm for underactuated surface vessels
with actuator saturation. The damping terms in the vehicle dynamics are unknown.
NN-based observer is developed to estimate the velocity of the vehicle with uncer-
tainties. Reference [30] developed a one-layer neural network tracking controller
with preprocessed input signals. The uniformly ultimately boundedness of the track-
ing error can be guaranteed with bounded external disturbances. In [31], multilayer
neural network technique was employed to deal with the problems of unmodeled
dynamics of underwater vehicle and environmental disturbances. According to bio-
inspired neural network architecture, Ref. [32] proposed a dynamic NN model to
plan a time-saving path in ocean current environments.
14 1 Introduction
Hidden
Input
Output
ηd η
AUV model
NN Controller
Fig. 1.9 Neural network control for AUV tracking
1.2.2.5 Sliding Mode Tracking Controller
Nonlinear dynamical systems suffer from the performance degradation caused by

uncertain conditions and disturbances. Sliding mode control (SMC) is one of the
most efficient control techniques aiming at eliminating the effect of bounded heavy
uncertainties or disturbances by means of a discontinuous controller. It uses a switch-
ing function as a key to prescribe a twofold behavior in the system. SMC has a variety
of attractive features. First, the controlled system with sliding mode exhibits strong
robustness properties with respect to both unavoidable parameter uncertainties and
external perturbations. Second, it is easy to implement in practice.
Reaching Phase
Sliding
Manifold
O
Sliding Phase
Fig. 1.10 Sliding mode control

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.
1.2.3 Cooperative Formation for Multiple AUVs
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
1.2.3.1 Centralized Formation Control
In centralized way, a control center on coast (or in a boat/vessel on water offshore)

collects all available data from AUVs and environment, as depicted by Fig. 1.11. In
order to enable AUVs keep a desired chape during the whole tracking procedure, a
centralized formation controller is then designed based on all available information.
With the reverse communication links, the mission decision on centralized formation
controller is transmitted to each AUV. Till now, many centralized formation con-
trollers have been developed for AUVs. For instance, a centralized leader–follower
formation controller was developed in [35] for the cooperative motion of multiple
AUVs with a vessel as the leader AUV. In [36], a centralized formation controller
was designed to pursue an underwater evader, where the relative position informa-
tion between central AUV and the other ones should be acquired for the formation
approach. Also of relevance, a joint target localization and formation scheme was
presented in [37] to maximize range-related performances for target localization and
formation control, in which the control center can calculate the control action for
each AUV. The main advantages of centralized formation controller include simple
design and convenient implementation in water. However, its disadvantages are also
obvious, e.g., the robustness and scalability are weak due to limited communication
ranges.
1.2.3.2 Distributed Formation Control
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-
Fig. 1.11 Description of the

centralized formation Target State Centralized Formation
controller Controller
AUV
State AUV 1 AUV n
Communication Channel
Controller 1 AUV 1
Target
State Controller 1 AUV 2
AUV
State
Controller n AUV n
Fig. 1.12 Description of the distributed formation controller
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.
1.2.3.3 Hierarchical Formation Control
The hierarchical way can be regarded as a compromise between centralized way

and distributed way, as depicted by Fig. 1.13. In some special scenarios, the pure
distributed fashion is not sufficient for AUVs to accomplish complicated tasks, due to
the limited sensing ability of AUVs. In this context, it becomes favorable for AUVs
to cooperatively operate, assisted by sub-control centers. Particularly, the centralized
formation controller sends its decision to subcontroller. The subcontroller analyses
the control signal, and hence, each subcontroller sends and feedback its state to the
communication channel. Repeating the above procedure, the whole formation shape
can be updated and maintained with the hierarchical formation controllers.
Recently, many hierarchical formation controllers have been developed for AUVs.
For instance, a virtual leader–follower hierarchical formation controller was devel-
oped in [43], where the formation shape can be ensured by hierarchical virtual struc-
ture. In [44], a finite-time formation controller with hierarchical leader–follower
structure was developed. More recently, logic-based approach was applied in [45]
for the formation of AUVs, where hierarchical three-level architecture of formation
system was conducted. Also of relevance, some other hierarchical formation con-
trollers can be found in [46, 47]. It should be emphasized that the advantages of
hierarchical formation controllers includes high scalability and reduced computing-
18 1 Introduction
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
Fig. 1.13 Description of the hierarchical formation controller
communicating complexity. Meanwhile, they also have the disadvantages of weak

robustness and scalability, with respect to the centralized way. However, the above
disadvantages can be overcome and omitted by optimized designs.
1.3 Research Challenges and Contributions of This Book
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.
• Asynchronous Clock. Propagation delay of underwater acoustic wave (≈ 1500

m/s) is five orders of magnitude higher than the radio wave (≈ 3 × 108 m/s).
Meanwhile, the global positioning system (GPS) is not available in water. Because
of the long propagation delay and the unavailability of GPS, the time clock of AUVs
is always asynchronous, i.e., the time clock is affected by unknown time skew and
offset. As a result, the clock synchronization-based range estimation strategy is
invalid for AUVs.
1.3 Research Challenges and Contributions of This Book 19
Table 1.1 Property comparison for the underwater communication mediums

Medium Range Rate Delay Disadvantage Advantage
(ms/km)
Acoustic ∼km ∼ kbps 667 Low Long range
bandwidth, distance
high delay,
high energy
consumption
Optical ∼100 m ∼ Mbps 0.03 Line of sight, High
short commu- bandwidth,
nication low delay, low
range energy cost
Radio ∼10 m ∼ Mbps 0.03 Short commu- High
nication bandwidth,
range low delay, low
energy cost
• Stratification Effect. Water medium is inhomogeneous, and hence, the propa-

gation delay of acoustic wave varies in different depth levels, which is called as
stratification effect. In view of the stratification effect, some researchers point out
that the isogradient sound speed profile can be employed to depict the acoustic
propagation speed. It should be mentioned that the ignorance of stratification effect
can increase the ranging biases and thereby reduce the localization accuracy of
AUVs.
• Node Mobility. Notice that the nodes in terrestrial Internet of things (IoTs) are
always static; however, AUVs can autonomously move due to the water current
and self motor. In view of this, the roundtrip propagation delays between any two
nodes are not equal, which make it hard to estimate the actual relative distance
between AUV and sensor node.
• Energy Constraint. Underwater nodes, including the sensor nodes and AUV, are
usually battery powered. It is difficult to replenish the batteries mounted AUVs
and sensor nodes. Besides that, the communication energy consumption in water
is much higher than the one in terrestrial environment. Although the energy can be
extended with added battery, the extension is at a cost of extra space occupation,
which will significantly increase the cost of AUVs or sensor nodes.
• Model Uncertainty. Typically, AUVs operate in harsh marine environment, where
their motions are strongly affected by the external disturbances such as tides,
currents, and streams. The internal model uncertainties and external disturbances
make it difficult or even impossible to establish an accurate model. Thereby, the
model uncertainty requires to be considered during the designs of tracking and
formation strategies.
With consideration of the above research challenges, this book is concerned on
the localization, tracking, and formation issues for AUVs. The structural relationship
of chapters can be depicted by Fig. 1.14. Particularly, Chap. 2 presents a ubiquitous
20 1 Introduction
Introduction (Chapter 1)
Localization Tracking Control Formation
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)
Future Research Directions (Chapter 9)
Fig. 1.14 Structural relationship of chapters in this book
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
1. Autonomous underwater vehicle (AUV) market. [Online]. Available: https://www.reportlinker.

com/p05916986/?utm_source=GNW. Accessed on May 2020
2. H. Huang, Y. Zheng, AoA assisted localization for underwater ad-hoc sensor networks, in
Proceedings of OCEANS 2016 MTS/IEEE Monterey, Monterey, Sept 2016, pp. 1–6
3. H. Huang, Y. Zheng, Node localization with AoA assistance in multi-hop underwater sensor
networks. Ad Hoc Netw. 78(1), 32–41 (2018)
4. Z. Gong, C. Li, F. Jiang, AUV-aided joint localization and time synchronization for underwater
acoustic sensor networks. IEEE Signal Process Lett. 25(4), 477–481 (2018)
5. J. Yi, D. Mirza, R. Kastner, C. Schurgers, P. Roberts, J. Jaffe, ToA-TS: time of arrival based
joint time synchronization and tracking for mobile underwater systems. Ad Hoc Netw. 34(1),
211–223 (2015)
6. B. Ferreira, A. Matos, N. Cruz, Optimal positioning of autonomous marine vehicles for under-
water acoustic source localization using TOA measurements, in Proceedings of 2013 IEEE
International Underwater Technology Symposium, Tokyo, Mar 2013, pp. 1–7
7. Q. Liang, B. Zhang, C. Zhao, Y. Pi, TDoA for passive localization: underwater versus terrestrial
environment. IEEE Trans. Parallel Distrib. Syst. 24(10), 2100–2108 (2013)
8. S. Poursheikhali, H. Zamiri-lafarian, TDOA based target localization in inhomogenous under-
water wireless sensor network, in Proceedings of 2015 ICCKE, Mashhad, Oct 2015, pp. 1–6
9. T. Wang, H. Xiong, H. Ding, L. Zheng, TDOA-based joint synchronization and localization
algorithm for asynchronous wireless sensor networks. IEEE Trans. Commun. 68(5), 3107–3124
(2020)
22 1 Introduction
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
To obtain the real-time position of AUV, many sonar-array-based localization meth-

ods have been proposed. Typically, the sonar arrays are towed by a ship or a sub-
marine, which makes them unsuitable for ubiquitous localization task. In addition,
if the ship or submarine that tows the sonar arrays is attacked by enemies, the entire
localization system will be paralyzed completely. Inspired by this, some researchers
employ inertial measurement unit (IMU) and Doppler velocity log (DVL) to locate the
position. However, IMU and DVL have the disadvantage of error accumulation since
the lack of necessary feedback mechanism, which in turn reduces the localization
accuracy. Fortunately, we notice that USNs can provide with us a promising solu-
tion. Especially, USNs consist of a large number of interconnected sensing units that
enable to perform collaborative monitoring tasks in a given water region. Compared
with sonar array system, USNs have the advantages of low cost, rapid deployment,
and self-organization. Meanwhile, the error accumulation in USNs can be shielded
via cooperation among different types of nodes.
Inspired by the advantages of USNs, some persistent localization schemes have
been proposed in the context of USNs. For example, Zhang et al. [1] studied the per-
sistent localization performances with respect to the underwater network topology,
and then the relationship between localization accuracy and network topology was
derived. In [2], an energy-efficient underwater target persistent localization scheme
was developed to prolong the network lifetime. Also of relevance, some other under-
water target localization and persistent localization schemes were developed in [1,
3–9]. However, these localization and persistent localization schemes rely on the
assumptions of accurate clock synchronization and straight-line data transmission.
However, they are strict for USNs, which have been pointed out in Chap. 1. Besides
that, the first-order linearization can introduce large model errors with the methods
in [5, 6]. To reduce model errors, the consensus-based unscented Kalman filter-
ing algorithm was developed in [10] to estimate static target. Recently, the artifi-
© 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:
1. Observer-based motion prediction. An UT-based localization estimator is

developed to estimate AUV. Subsequently, a terminal sliding mode velocity
observer is designed to predict the mobility of AUV. Compared with the posi-
tion acquisition strategies [24, 25], the UT-based localization estimator in this
chapter can relax linearization requirement and improve localization accuracy.
Different from [17–19], the proposed observer combines the position measure-
ment with AUV dynamic, through which the translational and angular velocities
of AUV can be accurately derived.
2.1 Introduction 27
2. Minimum rigid graph-based localization strategy. A minimum rigid graph-

based localization strategy is presented for sensor nodes to achieve persistent
localization. Compared with the existing works in [1, 8], the proposed localization
strategy in this chapter can balance the trade-off between localization accuracy
and energy consumption. In addition, it does not rely on the assumptions of clock
synchronization and straight-line transmission, which are different from the ones
in [3, 4, 8, 20].
2.2 Network Model and Problem Formulation
2.2.1 Network Model
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
η̇0 = Jv0 , Mv̇0 + Cv0 + Dv0 = τ , (2.1)

Fig. 2.1 Illustration of the network architecture
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
Mη̈0 + Cη̇0 + Dη̇0 = τ̄ , (2.2)
and its corresponding discrete-time form can be given as
(k + 1) = (k) + h((k))δ, (2.3)

−1
where = [ηT0 , (Jv0 )T ]T is the state vector, and h() = [Jv0 ; −M (CJv0 +
DJv0 − τ̄ )]. In addition, δ ∈ R+ is the local sampling period of AUV.
It is noted that the clock of AUV is not well synchronized with sensor nodes.
Based on this, the clock model of AUV is considered as follow, i.e.,
Tauv = αauv t + βauv , (2.4)
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
Fig. 2.2 Description of a ray

between sender S and
receiver R
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 θS = β0 − α0 and θR = β0 + α0 denote the ray angles at sender node S

and
receiver node R, respectively. Especially, α0 = arctan[(0.5ā
((xS − xR )2 + (yS − yR )2 ))/(b̄ + 0.5ā(zS + zR ))] is the
angle of the ray trajectory
deviates from this straight line. β0 = arctan[(zS − zR )/( (xS − xR )2 + (yS − yR )2 )]
is the angle of the straight line between node S and node R, with respect to the hor-
izontal axis. It is noted that, (xS , yS , zS ) denotes the position of sender node S and
(xR , yR , zR ) denotes the position of receiver node R.
The energy consumptions to transmit and receive b bits packet over distance d
can be formulated as [30]
Etx (b, d ) = bPr Tp d λ g(f )d ,

(2.6)
Erx (b, d ) = bPr Tp ,
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.
2.2.2 Problem Formulation
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).
Problem 2.2 (Minimum Rigid Graph-Based Localization) To realize ubiquitous

localization for AUV, sensor nodes in USNs require to keep a high network connec-
tivity. Nevertheless, the network scale of USNs is large, and more importantly, the
energy consumption of underwater acoustic communication is much higher than the
one with terrestrial radio wave. Hence, it is difficult to balance the trade-off between
localization accuracy and energy consumption. To handle this issue, a rigid graph-
based localization approach is designed to achieve persistent localization of AUV,
through which a minimum rigid graph and unscented Kalman filtering (UKF) algo-
rithm are jointly employed to improve energy efficiency and localization accuracy.
2.3 Main Results
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.
Fig. 2.3 Framework of the rigid graph-based localization approach

2.3 Main Results 31
2.3.1 Observer-Based Motion Prediction for AUV
In this section, an UT-based localization estimator is employed by sensor nodes to

obtain the initial position of AUV, i.e., p0 (0) = [x0 (0), y0 (0), z0 (0)]T , and then a
terminal sliding mode velocity observer is designed to predict the trajectory of AUV,
i.e., η0 (k) = [x0 (k), y0 (k), z0 (k), ϕ0 (k)]T for k ≥ 1. To locate p0 (0), the following
steps are conducted.
Step 0. Initialization
When AUV enters into the detection range of sensor nodes, the neighboring sensor
nodes are selected as the active sensor nodes. They are indexed by 1, . . . , N1 , where
N1 denotes the total number. Besides that, the position of active sensor node i is
expressed as pi = [xi , yi , zi ]T for i ∈ {1, . . . , N1 }, and its detection range is Ra . In
addition, it is assumed that the model parameters of AUV are preknown to all sensor
nodes.
Step 1. Data Collection
AUV broadcasts an initiator message to active sensor nodes. To be specific, the
initiator message includes its sending time Ta,a , its initial yaw angle ϕ0 (0), and the
local sampling period δ. At time ta,i , active sensor node i ∈ {1, 2, . . . , N1 } receives
the initiator message from AUV. Then, active sensor node j sends its message to
active sensor node i at time tj,j , where i ∈ {2, . . . , N1 } and j ∈ {1, . . . , i − 1}. Active
sensor node i receives this message at time tj,i . Accordingly, the message sent by the
active sensor node j includes pj , ta,j , tj,i , and tj,j . After active sensor node, N1 receives
message from initial active sensor node N1 − 1, a round of localization procedure
is finished. Based on this, the collected data by active sensor node N1 includes Ta,a ,
N1 ,i−1
ϕ0 (0), δ, {tj,i , tj,j , pj }i=2,j=1 , and {ta,i }Ni=1
1
.
Step 2. Construction of Time Differences
In order to eliminate the effect of clock parameters (i.e., αauv and βauv ), the time
difference can be constructed as
Ti,j = ta,i − ta,j , (2.7)
where i ∈ {2, . . . , N1 } and j ∈ {1, . . . , i − 1}.

It is assumed that each node has the same measurement quality, i.e., the noise of
local measurement is a random variable with zero mean and variance σmea 2
. In view of
(2.7), the relationship between time difference and propagation delays can be given
as
Ta,a − βauv Ta,a − βauv
Ti,j = + τa,i + a,i − + τa,j + a,j
αauv αauv (2.8)
= τa,i − τa,j + i,j ,
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
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
Step 3. Optimization Problem Establishment

Combining (2.7) and (2.8) into vector-matrix form, one has
φ = ξ (p0 (0), pi ) + w, (2.9)
where φ = [T2,1 , . . . , TN1 ,N1 −1 ]T denotes measurement vector, ξ (p0 (0), pi ) =

[τa,2 − τa,1 , . . . , τa,N1 − τa,N1 −1 ]T and w = [ 2,1 , . . . , N1 ,N1 −1 ]T denote noise vec-
tor with covariance R.
Therefore, the optimization problem can be formulated as
p̃0 (0) = argmin[(φ − ξ (p0 (0), pi ))T (φ − ξ (p0 (0), pi ))]. (2.10)
p0 (0)
Step 4. Solution to the Optimization Problem

Denote the lth iteration of p0 (0) as p̃l0 (0) = [x̃0l (0), ỹ0l (0), z̃0l (0)]T . The linearization
of ξ (p0 (0), pi ) can be given as
ξ (p0 (0), pi ) = ξ (p̃l0 (0), pi ) + H(p̃l0 (0)) l

, (2.11)
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:—
“Without reference, however, to these extreme cases, it must be repeated,

that the use of water, simply as a diluent, scarcely receives attention and
discrimination enough in our English practice.”
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.”
With this quotation we take leave of Sir Henry Holland, merely

observing, that no Hydropathist could say more on the subject than
he has done, and that the Continental practice referred to, of drinking
large quantities of water, conjoined with ample exercise and regular
habits of life, is precisely that practice which Hydropathy enjoins.
Sir John Forbes, a physician already quoted, says, on water drinking

—
“The water cure is a stomachic, since it invariably increases the appetite.”

Dr. Pereira states—
“It is a vital stimulus, and is more essential to our existence than aliment.”
Liebig, the celebrated physiological chemist, bears similar testimony,

viz.—“It increases the appetite.”
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?
Some objectors say, “water drinking thins the blood.” After

demolishing these objections by arguments which we regret we have
not space to quote, Dr. Gully concludes his observations as follows;
—
“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.’ ”
It is a curious fact that in all the medical works which treat of

anaemia, or bloodlessness, “allusion is never once made to water-
drinking as a known cause—not even to the possibility of its being a
cause of it.”
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.
Perhaps there is no disease which would appear, at first sight, so

little suited for Hydropathic treatment as cholera; 15 that disease for
the successful treatment of which we have been hitherto
accustomed to consider stimulants and hot applications of all kinds
as indispensably necessary, and yet there is no disease, in the
treatment of which Hydropathy has been more successful.
The principles of its treatment, by the water system, are so sensibly

and rationally put forward in the pamphlet entitled, “An Address, &c.,”
that, as we think, the greatest sceptic must be convinced of the truth
of the doctrines it propounds, we strongly recommend its perusal to
our readers. Of the many cases treated by the author, all, we are
told, recovered, whilst not a single instance of secondary fever—the
invariable accompaniment of the Allopathic treatment, and only
secondary in danger to the disease itself—occurred. The necessary
prevalence of this secondary fever in the one case, and its
[31]absence in the other, are beautifully explained, on natural
principles, at pages 9 and 10. Though the pamphlet in question is
anonymous, and the author has taken some pains to explain his
reason for concealing his name, yet he has unwittingly betrayed his
identity in the following extract from a letter from Lieut.-Colonel
Cummins, c.m., who, having tried the system as an amateur, in
America, thus writes of it:—
“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.
“I never saw cholera of so frightful a character; that at Quebec, which you

recollect was so near doing for me, was nothing to it; the violence of the
spasms was such that blood oozed out through all the pores of the skin,
especially with the niggers. It did not give the slightest warning; the men
often fell while at work, and before four hours were dead.”
The following statement, extracted from a letter written by Mr. James

Morgan of Cork, and which appeared in the Limerick Chronicle, 4th
April, 1849, affords a remarkable instance of the beneficial effects of
fresh air and cold water, so strongly insisted upon by Dr. Barter, and
corroborating the practice which, on theoretical grounds, he
recommends:
“In a temporary cholera hospital at Gloucester, there were sixteen

patients, one of whom was an interesting young female, between fifteen
and sixteen years of age, for whose recovery the attending physician (Dr.
Shute) was most anxious. On leaving the hospital in the evening, the girl
was in collapse, and quite blue; he called the nursetender, and bade her
be attentive to her, and give her whatever she may call for, as all hopes
had vanished. In the course of the night the nurse went to increase the fire
which was near the girl’s berth in the ward; but she begged the woman not
to do so, as she was almost suffocated, and, at the same time, asked for a
drink. The nurse brought her a bowl of tea, which was rejected, but she
requested water. Remembering the doctor’s directions, the nurse, not
without some reluctance and apprehension, brought her a pint mug full of
water, which she drank with avidity; and continued to call for water about
every five minutes, until she had taken two gallons of it; when she fell into
a profound sleep, in which she was found by the doctor in the morning,
when her natural complexion reappeared, and she was, to his
astonishment, in a state of convalescence. Having with amazement
elevated his eyes, exclaiming, ‘this is something like a miracle!’ he called
the nursetender, who related what had taken place; and perceiving the
window open over the patient’s berth, he asked why it was not shut? and
was told by the attendant, that it was left open at the earnest desire of the
girl. The doctor immediately ordered all the windows of the ward to be
opened, the heavy bed covering on the patients to be removed, and
replaced by light rugs; directed that no drink should be given but cold
water, and the result was, that the whole sixteen persons were cured of
cholera; one, however, died of consecutive fever, produced by eating too
much chicken and drinking too much broth whilst convalescent. The case
[32]was reported to the Government Board of Health, then sitting in
London; and similar treatment was pursued by all the medical men in and
about Gloucester with the most complete success. The report, names of
the doctors, and all the correspondence are minutely detailed in the
columns of the Chronicle in the year 1832.
“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.” * * * * *
The following extract is taken from Braithnorth’s “Retrospect of

Medicine,” a standard professional work:—
“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.”
We make no comments on the foregoing, leaving the public to draw

their own conclusions from the facts stated. In setting the facts
before them, we feel we have done our duty; we leave the leaven to
work in their minds, and produce its own result on their future
conduct.
In condemning the mistaken administration of hot stimulants, such

as “hot punch,” &c., Dr. Barter proceeds:—
“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.
Sir Bulwer Lytton thus sums up his impressions of Hydropathy:—
“Those cases in which the water-cure seems an absolute panacea, and in

which the patient may commence with the most sanguine hopes, are—
first, rheumatism, however prolonged, however complicated. In this the
cure is usually rapid—nearly always permanent. 16 Secondly, gout: here its
efficacy is little less startling to appearance than in the former case; it
seems to take up the disease by the roots; it extracts the peculiar acid
which often appears in discolorations upon the sheets used in the
application, or is ejected in other modes. But here, judging always from
cases [34]subjected to my personal knowledge, I have not seen instances
to justify the assertion that returns of the disease do not occur. The
predisposition—the tendency, has appeared to remain; the patient is liable
to relapses, but I have invariably found them far less frequent, less
lengthened, and readily susceptible of simple and speedy cure, especially
if the habits remain temperate.”
If it be asked why Hydropathy has proved itself so effective a remedy
in curing rheumatism, we would answer, on account of its great
power in strengthening and invigorating the stomach and digestive
organs, in the derangement of which, the cause of that disease is to
be found. Rheumatism proceeds from a sluggish circulation in the
extremities, consequent on a low vitality in the system, arising from a
derangement of the digestive organs and viscera; if these latter were
sound and free from irritation, all the cold and wet, we could possibly
be exposed to, would fail to produce that inflammation of the sheaths
of the muscles in which rheumatism consists. That Hydropathy is
capable of strengthening and invigorating these organs, is well
known to all who have tried it, and is even admitted by its greatest
opponents when they state, “Oh! it is good for the general health,” for
it is utterly impossible for the “general health” to be good without a
sound digestion.
With respect to gout, a permanent cure from it is rarely to be found,

and why?—Because few people have either the time or patience to
continue long enough under treatment for its total eradication,
running away from an “establishment” the moment they get relief
from the pressing fit, and consequently the disease recurs. Now, of
all diseases, gout is perhaps the most tedious of permanent cure,
the visceral irritation which gives rise to it being always inveterate
and of long duration, and nothing short of chronic treatment—
treatment continuing for years instead of months, will remove it. Dr.
Gully observes respecting it:—
“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.”
On the effects of colchicum he, further on, observes:—
“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.’ ”
When the drugging practitioner drives the inflammation from the

extremities to a more dangerous internal position, he congratulates
himself on having cured the gout; but what in reality has he done?—
By his mischievous interference with nature, he has endangered his
patient’s life and shaken his constitution; whilst the gouty irritation,
which causes the complaint, remains unsubdued, ready to be
transferred at a moment to the head or heart, the practitioner having
cleverly banished it from its original harmless position. It is in this
way also that the Allopathist cures skin diseases, driving in the
irritation which nature is struggling to drive out; this he eventually
succeeds in doing, by weakening the powers of the system, and then
fancies the disease is cured, whilst the patient pays in the long run
for these hostile operations against nature.
But we have interrupted Sir Bulwer Lytton,—he thus proceeds:—

“Thirdly, that wide and grisly family of affliction classed under the common
name of dyspepsia. All derangements of the digestive organs, imperfect
powers of nutrition—the malaise of an injured stomach, appear precisely
the complaints on which the system takes firmest hold, and in which it
effects those cures that convert existence from a burden into a blessing.
“Hence it follows that many nameless and countless complaints,

proceeding from derangement of the digestive organs, cease as that great
machine is restored to order. I have seen disorders of the heart which
have been pronounced organic by no inferior authorities of the profession,
disappear in an incredibly short time; cases of incipient consumption, in
which the seat is in the nutritious powers; hæmorrhages, and various
congestions, shortness of breath, habitual fainting fits, many of what are
called improperly nervous complaints, but which in reality are radiations
from the main ganglionic spring: the disorders produced by the abuse of
powerful medicines, especially mercury and iodine; the loss of appetite,
the dulled sense and the shaking hand of intemperance, skin complaints,
and the dire scourge of scrofula;—all these seem to obtain from
Hydropathy relief,—nay, absolute and unqualified cure, beyond not only
the means of the most skilful practitioner, but the hopes of the most
sanguine patient.”
[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.
We wonder how long the public will continue to poison 18 their

systems with mercury, colchicum, iodine, and prussic acid, because
a physician chooses to tell them that a mode of treatment which he
has never investigated “will kill them.”
It may not be uninteresting to observe, that under Hydropathic

treatment, chronic disease frequently becomes acute; for, as the
body improves in strength, the more acutely will any [37]existing
disease develop itself, and for the following reason: pain is caused
by an effort of nature to relieve the system of some morbid influence
residing in it, and the stronger the constitution, the greater efforts will
it make to remove that morbid influence, and therefore the greater
will be the pain; but on the other hand, when the body is enfeebled,
its efforts to relieve itself, though continual, are weak and inefficient,
and the disease remaining in the system, assumes the chronic and
less painful form. Now with these facts before them, we have been
amused at hearing physicians observe, in their efforts to decry the
“Water System,” “Oh it is good for the general health, but nothing
more,” a result albeit, which unfortunately the Allopathic system
cannot lay claim to. When speaking thus they do not however reflect
that they are affording the strongest possible testimony in support of
the system which they seek to decry, inasmuch as every
physiologist, from Cape Clear to the Giant’s Causeway, admits the
principle, that the cure of disease is to be sought for in the powers of
the living organism alone; and it must be evident that the more you
strengthen that organism, the more you increase its powers to cure
itself, and diminish its liability to future disease.
Having trespassed thus far on the attention of our readers, we would

conclude by inviting them and the medical profession, generally, to a
calm and dispassionate investigation, as far as the opportunities of
each allow, of the relative merits of the Allopathic and Hydropathic
systems, approaching the investigation, as far as possible, with a
mind devoid of prejudice and bigotry. Their duty to themselves and to
society demands this inquiry at their hands—two antagonistic (we
use the term advisedly) systems are presented for their acceptance
—which will they lay hold of? To assist them in determining this point
we would recommend for their quiet perusal either or all of the works
alluded to in this article, 19 the study of which will be found interesting
and profitable. If they conclude that drugs are wholesome, let them
by all means be swallowed; but if they are proved to be injurious,
deleterious and unnecessary, then away with them;—if opiates are
innocuous let them be retained, but if they congest the liver, sicken
[38]the stomach, and paralyze the actions of the vital organs, the
sooner they are erased for ever from the Hygienic Pharmacopeia the
better—let them gracefully retire before the improved system of hot
stupes, fomentations, and the abdominal compress.
The very simplicity of the processes of the “water-cure,” which

people cannot believe capable of producing the effects ascribed to
them, combined with a belief, ingrained by long habit, in the absolute
necessity for drugs in curing disease, have chiefly militated against a
more extended reception of Hydropathy by the lay public; but when
they reflect that all the powers of the medical art range themselves
under two great categories, stimulants and sedatives—blistering,
bleeding, drugs, and leeching—acknowledging no other objects, they
cannot but admit the possibility of Hydropathy possessing the
powers attributed to it, since its bracing and soothing properties
cannot be questioned. Were, however, the position of affairs
reversed, and Hydropathy become as old a system as the Allopathic
this belief, in the efficacy of the old school might be securely
entertained; for no one would think for a moment of exchanging a
system, fixed, intelligible, and certain in its action, as based on
scientific principles, and consonant with the laws of physiology, for
the uncertain, groping, empirical, and injurious practice of drug
medication.
We would ask the medical profession of Ireland to reflect on the fact,

that Dr. Barter’s establishment at Blarney, contains at this moment
upwards of 120 patients, with many more frequently seeking for
admission within its walls, most of whom leave the establishment
ardent converts to Hydropathy, determined for the rest of their lives
to “throw physic to the dogs,” fleeing from it as from some poisonous
thing. It will not do for them to “pooh-pooh” the system, and tell their
patients, as many of them do, that it will kill them; 20 [39]such language
only betrays ignorance on their part, and will not put down a system
which daily gives the lie to their predictions by affording ocular
demonstration of its efficacy, in the restored health and blooming
cheek of many a once emaciated friend. Men are too sensible now-
a-days to pin their faith on the dictum of a medical man, who runs
down a system without fairly investigating it, and examining the
principles on which it acts, to say nothing of the prejudice he must
feel in favour of his own particular system; but if a mode of treatment
be rational, producing cures where every other system of treatment
has failed, and recommend itself to the common sense and reason
of mankind, we believe that such a principle will make its way
despite of the opposition of all the physicians that ever lived; and this
very progress the water cure is now making.
We would in conclusion apostrophize Hydropathy, in the words of the

American traveller, who gave vent to his feelings on first beholding
the falls of Niagara, by exclaiming, “Well done, Water!!”
1 1. The Water Cure in Chronic Disease. By James M. Gully, M.D. London:

Churchill.
2. The Water Cure. By James Wilson, M.D. London: Trubner and Co.
3. Hydropathy. By Ed. Wm. Lane, M.D. London: Churchill.
4. Confessions of a Water Patient. By Sir E. Bulwer Lytton, Bart, M.P. London: H.
Baillière.
5. An Address to the Public on the Prevention and Treatment of Cholera on
rational principles. Cork: Geo. Purcell and Co.
6. A few Facts forgotten by the Faculty. By S. B. Birch, M.D. London: H. Baillière. ↑
2 Because either they disapprove of whatever is not practiced by themselves, or
they are ashamed to follow new opinions, and to acknowledge when old, the
errors they had imbibed in youth. ↑
3 The ganglionic nerves are those which cover the stomach, and regulate the
digestive organs: they are also called the “Solar Plexus.” ↑
4 The late melancholy case of Mr. Stafford O’Brien is an instance of this injurious
practice; that gentleman was copiously bled, doubtless that he might be the
better enabled, in his so enfeebled condition, to resist the action of a powerful
poison (opium) afterwards administered with deadly effect. We cast no imputation
whatever on the attendant physician in this case, as we believe the treatment
pursued by him was strictly that enjoined by the orthodox school. Yet, if one
wished to destroy life, could they take a surer means of doing it? ↑
5 “The generality of men are not so much accustomed to pursue this or that
course, in consequence of their previous conviction that it is right, as to believe
that it is right, because they have been accustomed to pursue it.”—Archbishop
Whately. ↑
6 We do not pretend to assert, that consumption is curable when organic
disease of the lungs has actually been established, but we maintain that the
disease is perfectly curable in its incipient stages, though not by drugs nor
banishment to a foreign clime. The latter may somewhat prolong the disease, but
will not cure it, unless by accident, when of a very mild form. ↑
7 The temperature and vitality of our bodies depend upon the continued and rapid
combination of oxygen with the oxidizable products of the blood; if the
necessary supply of oxygen be interfered with, the vitality of the system flags, and
disease results. ↑
8 The very name of scrofula points to the origin of the disease, it being derived
from the Latin Scrofa, a pig (quod sues præcipue hoc morbo vexantur. Cels. 5,
38), in allusion to the condition of the skin in those persons in whom a scrofulous
habit has been engendered. It has been proved beyond contradiction that the
partial closure of the pores, which every one suffers from in some degree, is the
chief source of scrofula in all its hideous forms. ↑
9 When blood is overloaded with carbon, and deprived of its necessary supply of
oxygen, the term “veinous” is applied to it. ↑
10 Where consumption has been relieved by residence abroad, the benefit derived
must be attributed to the action on the skin produced by the hot climates to
which the patient is usually ordered, but recovery in this way has been confined to
very mild forms of the disease, and cannot be looked upon as a scientific mode of
treatment; the improved action of the skin deserving to be considered rather as
induced accidentally than by design; as otherwise more attention would have been
paid to so important a matter, and there would have been no necessity for ordering
the patient abroad, as similar results could have been obtained much more easily
and effectually by keeping him at home; the use of the Turkish bath conferring all
the benefits of increased temperature, followed by the tonic effects of cool air and
water, by which the debilitating effects of continual residence in a warm climate are
obviated. ↑
11Dr. Hufeland remarks—“The more active and open the skin is the more secure
will the people be against obstructions and diseases of the lungs, intestines,
and lower stomach; and the less tendency will they have to gastric (bilious) fevers,
hypochondriasis, gout, asthma, catarrh, and varicose veins.” ↑
12The wearing of flannel close to the skin has a two-fold injurious effect:—First, by
driving the blood from the surface, whereby the activity of the skin is seriously
impaired; and secondly, by shutting out the air, and so preventing it from having
access to the blood, to aerate and purify it. ↑
13 By healthy waste, we mean waste accompanied by corresponding renewal. ↑
We have seen consumptive patients arrive at Blarney shivering with cold though
14
swathed in flannels, who before leaving it were able to wear clothing in winter,
under which they previously would have shivered in the hottest day of summer. ↑
15 The great mortality which has attended the Allopathic treatment of cholera,
ought to make us have little compunction in trying something new. There is no
fear, in this case, of our “jumping from the frying-pan into the fire;” we are already
in it—let us quench it. ↑
16 Dr. Russell, a well-known Homœopathic author, appears to give the palm to
Hydropathy in some rheumatic cases. He thus writes: “In regard to rheumatism,
I am inclined to think that there are some varieties of this complaint which utterly
defy all Homœopathic medicines, from the deeply morbid condition of the blood;
and that in these cases a thorough water course, by effecting a rapid and total
renovation of this fluid, might enable our remedies to act more beneficially.”
Contrast this liberality with that of the Allopathic physician. ↑
17 We were told by our physician that the water system would kill us, as we had
“not sufficient reaction to stand it.” Had he, however, understood anything of its
working, he could not have made this observation, as Hydropathy implies cold
water only in those cases to which cold water is suited; and if he had asked Dr.
Barter, he would have told him that the chief thing he had to guard against in
practice, was excessive reaction, instead of the want of it. We ran the gauntlet,
however, and can truly say we never knew what real health was until we did so,
and forswore the use of drugs. ↑
18 The administering of poison to cure disease, is nothing short of a contradiction
of terms; for the word poison, if it means anything, means something injurious
to bodily health, and therefore incompatible with its welfare. We might as well try to
strengthen a man by bleeding him. ↑
19To those who desire a detailed and scientific account of the water cure in a
popular form, we recommend “The Confessions of a Water Patient,” by Sir
Bulwer Lytton; “The Water Cure,” by Dr. Wilson; and “Hydropathy,” by Dr. Lane;
whilst to those desiring a learned, lucid, and most able scientific treatise, we would
suggest “The Water Cure in Chronic Disease,” by Dr. Gully; and “Domestic
Hydropathy,” by Dr. Johnson. ↑
20 A friend of ours was told by a physician in whom many place confidence, that if
he opted to take the Turkish bath, it would kill him. Having, however, read
something on the subject, he went to Blarney, tried the bath, luxuriated in it, and
derived the greatest benefit from its use. We can tell the reader that this physician
at the time he prophesied, had never visited the Blarney bath, nor could he have
known any thing about it, as no description of it had been published at the time.
Under such circumstances an unsophisticated mind would think it more becoming
for him to have said—“I cannot advise you in this matter, as I have not studied the
subject; what you do must be on your own responsibility.” It is such illiberality of
mind as the above—such a want of pursuing truth for its own sake—which has
brought discredit on the medical profession, and loosened its hold on the public. ↑
[Contents]
THE TURKISH BATH. 1
“Come hither, ye that press your beds of down

And sleep not, see him sweating o’er his bread
Before he eats it.—’Tis the primal curse,
But softened into mercy, made the pledge
Of cheerful days, and nights without a groan.”—Cowper.
“Melancholy is overcome by a free perspiration; and cheerfulness, without

any evident cause, proceeds from it.”—Sanctorius.
What is a Turkish bath? Should this question be asked by any of our

readers, we would answer, that it is a bath differing from all other hot
baths in this important particular, viz., that the heated medium is air
instead of water; and that all parts of the body, when in the bath, are
subjected to an even and equal temperature. The result of which is,
that inasmuch as man was constituted to breathe air instead of
vapour, the Turkish bath may be enjoyed for hours at a time, without
[40]inconvenience; whereas in the vapour-bath the patient is unable
to remain in it for more than about a quarter of an hour, in
consequence of a feeling of suffocation, from want of a sufficient
supply of air to the lungs. And further, there is this difference
between the two baths, that in the case both of the vapour-bath and
the vapour-box 2 the pulse is materially raised, whilst in the Turkish
bath the pulse seldom rises above its normal state, which shows that
the circulation is very little affected by it—an all-important fact, which
is thus accounted for:—The normal temperature of the human body,
when in a state of health, is about 98° Faht., a temperature which
cannot be much augmented or diminished without producing
injurious results in the system; but as it is impossible always to
maintain so low temperature about us, Nature has provided, by
means of perspiration, a safety valve, by which the human body is
protected from the evil consequences which would arise from its
exposure to a high temperature—the principle on which she acts
being as follows:—It is a physical law that whenever evaporation
takes place a considerable amount of latent heat, (i.e. heat not
sensible to the thermometer), is absorbed, by which abstraction of
heat the temperature of the body from which the evaporation
proceeds is greatly lowered; but as evaporation consists in the
absorption of vapour by the surrounding air, it is evident that no
evaporation can take place where that air is already saturated with
moisture, and it is also evident that the amount of evaporation will
depend on the dryness of that air. Accordingly, in the Turkish bath,
the air being almost dry, when perspiration takes place it is followed
by a rapid evaporation which cools the body, and prevents its
temperature from rising above a healthful limit; whereas, in the
vapour-bath and vapour-box, the air being saturated, with moisture,
evaporation cannot take place, and consequently as no means for
reducing the high temperature of the body exist, the heat is thrown in
upon the system, raising the pulse, producing feverish headache,
and other symptoms of a highly deranged circulation; whilst a further
derangement arises from an insufficient supply of air to the surface
of the [41]body. In the Turkish bath, again, the system, feeling that it
has an ample supply of air, is not called upon to quicken the
circulation through the lungs in order to obtain an increased supply,
and thus another source of feverish excitement is obviated. These
and other considerations give the Turkish bath the pre-eminence,
longo intevallo, over all other artificial modes 3 yet invented for acting
on the skin by perspiration.
It may be observed that, cæteris paribus, the strength of each

person’s constitution is directly proportional to the quantity of oxygen
which his system is capable of imbibing, for on this the vitality of his
system and the purity of his blood, and therefore his health,

Ebook Autonomous Underwater Vehicles Localization Tracking and Formation Cognitive Intelligence and Robotics Yan Online PDF All Chapter

Uploaded by

Copyright:

Available Formats

You might also like

Ebook Autonomous Underwater Vehicles Localization Tracking and Formation Cognitive Intelligence and Robotics Yan Online PDF All Chapter

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Ebook Autonomous Underwater Vehicles Localization Tracking and Formation Cognitive Intelligence and Robotics Yan Online PDF All Chapter

Uploaded by

Copyright:

Available Formats

Autonomous Underwater Vehicles:

Localization, Tracking, and Formation

Autonomous Underwater Vehicles Jing Yan Xian Yang

Good Robot, Bad Robot: Dark And Creepy Sides Of

Embedded Robotics: From Mobile Robots to Autonomous

Cambridge IGCSE and O Level History Workbook 2C - Depth

Received Signal Strength Based Target Localization and

Autonomous Driving and Advanced Driver-Assistance

Autonomous and Connected Vehicles Network Architectures

Autonomous Vehicles Volume 1 Using Machine Intelligence

Jing Yan · Xian Yang · Haiyan Zhao ·

More information about this series at http://www.springer.com/series/15488

ISSN 2520-1956 ISSN 2520-1964 (electronic)

The results presented in this book mark a substantial contribution to nonlinear

leader–follower strategy, tracking and formation controllers are developed to stabilize

Qinhuangdao, China Jing Yan

3.3.2 Model-Free Tracking Controller Design . . . . . . . . . . . . . . . . . 70

6.5 Simulation and Experimental Studies . . . . . . . . . . . . . . . . . . . . . . . . . . 149

R Field of real numbers

λ(A) Eigenvalue of matrix A

UAVs Unmanned aerial vehicles

1.1 Underwater Observation Network

Fig. 1.1 Architecture of underwater observation network via AUVs

are cooperatively assisted to support the decision-making process of AUVs. Based

1.2 An Overview of the Existing Works

1.2.1 Ubiquitous Localization of AUVs

According to the different measurement ways, the ubiquitous localization schemes

1.2.1.1 AOA-Based Localization

The positions of AUVs can be estimated by AOA measurement. Particularly, each

Fig. 1.2 An illustration of

According to Stanfield’s method, it is assumed that the measurement noise is

1 sin2 (βi − θi (x))

1 [(y − yi ) cos βi − (x − xi ) sin βi ]2

(x̂, ŷ) = (AT R−1 S−1 A)−1 AT R−1 S−1 b, (1.5)

1.2.1.2 TOA-Based Localization

Fig. 1.3 An illustration of the TOA measurement

Ti = (ti,a − ta,a ) − τi ,

Based on the principle of trilateration, the localization optimization problem can

1.2.1.3 TDOA-Based Localization

Fig. 1.4 An illustration of

t̃i,j = ti,j + ni,j , (1.10)

ϕ(x) ≈ ϕ(x0 ) + ∇ϕ(x0 )(x − x0 ), (1.13)

1.2.1.4 RSS-Based Localization

Fig. 1.5 An illustration of

where f is the frequency of acoustic communication.

Based on this, the localization optimization problem of AUV is given as

x̂ = arg max p(P; x). (1.18)

It is noted that the optimization problem in (1.18) is highly non-convex. Thus, it

1.2.2 Persistent Tracking Control for a Single AUV

1.2.2.1 PID Tracking Controller

e(t) = ηd (t) − η(t). (1.20)

Fig. 1.6 PID control for AUV tracking

1.2.2.2 Fuzzy Tracking Controller

Fig. 1.7 Fuzzy control for AUV tracking

tive fuzzy-based controller was designed to guarantee tracking performance, where

1.2.2.3 Adaptive Tracking Controller

Fig. 1.8 Adaptive control

Adaptive control, nevertheless, as a branch of system theory, is curtailed in indus-

Ti = (ti,a − ta,a ) − τi ,

t̃i,j = ti,j + ni,j , (1.10)