Communications for Control in Cyber Physical Systems: Theory, Design and Applications in Smart Grids

Communications for Control in Cyber Physical Systems

Theory, Design and Applications in Smart Grids

First Edition

Husheng Li

Table of Contents

Cover image

Title page

Copyright

Dedication

Biography

Preface

Chapter 1: Introduction to cyber physical systems

Abstract

1.1 Introduction

1.2 Elements of a CPS

1.3 What is included and what is missing

Chapter 2: Basics of communications

Abstract

2.1 Introduction

2.2 Information measures

2.3 Communication channels

2.4 Source coding

2.5 Modulation and coding

2.6 Networking

2.7 Typical communication systems

2.8 Conclusions

Chapter 3: Basics of control

Abstract

3.1 Introduction

3.2 Modeling of controlled dynamical systems

3.3 Observability and controllability

3.4 Optimal control

3.5 Conclusions

Chapter 4: Typical cyber physical systems

Abstract

4.1 Introduction

4.2 Power networks

4.3 Robotic networks

4.4 Conclusions

Chapter 5: Communication capacity requirements

Abstract

5.1 Introduction

5.2 Deterministic system: stability

5.3 Stochastic systems: estimation

5.4 Stochastic systems: stability

5.5 Stochastic systems: reduction of shannon entropy

5.6 Networked stochastic systems

5.7 Control communication complexity

5.8 Control and information in physics

5.9 Conclusions

Chapter 6: Network topology design

Abstract

6.1 Introduction

6.2 WDM networks and design constraints

6.3 Optimization based on topology design

6.4 Team decision theory

6.5 Conclusions

Chapter 7: Communication network operation for CPSs

Abstract

7.1 Introduction

7.2 Hybrid system modeling for CPSs

7.3 Optimization of scheduling policy

7.4 Scheduling: other approaches

7.5 Routing

7.6 Conclusions

Chapter 8: Physical layer design

Abstract

8.1 Introduction

8.2 Adaptive modulation

8.3 Source coding in a CPS: Point-to-point case

8.4 Source coding in a CPS: Distributed case

8.5 Physical dynamics-aware channel decoding

8.6 Control-oriented channel coding

8.7 Channel coding for interactive communication in computing

8.8 Conclusions

Bibliography

Index

Copyright

Morgan Kaufmann is an imprint of Elsevier

50 Hampshire Street, 5th Floor, Cambridge, MA 02139, USA

© 2016 Elsevier Inc. All rights reserved.

No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher. Details on how to seek permission, further information about the Publisher’s permissions policies and our arrangements with organizations such as the Copyright Clearance Center and the Copyright Licensing Agency, can be found at our website: www.elsevier.com/permissions.

This book and the individual contributions contained in it are protected under copyright by the Publisher (other than as may be noted herein).

Notices

Knowledge and best practice in this field are constantly changing. As new research and experience broaden our understanding, changes in research methods, professional practices, or medical treatment may become necessary.

Practitioners and researchers must always rely on their own experience and knowledge in evaluating and using any information, methods, compounds, or experiments described herein. In using such information or methods they should be mindful of their own safety and the safety of others, including parties for whom they have a professional responsibility.

To the fullest extent of the law, neither the Publisher nor the authors, contributors, or editors, assume any liability for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein.

Library of Congress Cataloging-in-Publication Data

A catalog record for this book is available from the Library of Congress

British Library Cataloguing-in-Publication Data

A catalogue record for this book is available from the British Library

ISBN: 978-0-12-801950-4

For information on all Morgan Kaufmann publications visit our website at https://www.elsevier.com/

Publisher: Todd Green

Acquisition Editor: Brian Romer

Editorial Project Manager: Amy Invernizzi

Production Project Manager: Mohana Natarajan

Cover Designer: Vicky Pearson Esser

Dedication

To my dear wife, Min Duan, and my lovely children, Siyi Li and Mohan Li.

Biography

Husheng Li received BS and MS degrees in electronic engineering from Tsinghua University, Beijing, China, in 1998 and 2000, respectively, and a PhD degree in electrical engineering from Princeton University, Princeton, NJ, in 2005. From 2005 to 2007, he worked as a senior engineer at Qualcomm Inc., San Diego, CA. In 2007, he joined the EECS Department of the University of Tennessee, Knoxville, TN, as an assistant professor. He became an associate professor in 2013. His research is mainly focused on wireless communications, cyber physical systems, networking, statistical signal processing, smart grids, and game theory. He has received the Best Paper Awards from EURASIP Journal of Wireless Communications and Networks 2005, EURASIP Journal of Advances in Signal Processing 2015, IEEE ICC 2011, and IEEE SmartGridComm 2012 and the Best Demo Award of IEEE Globecom 2010.

Preface

Husheng Li

It is prohibitively difficult to introduce all the aspects of a cyber physical system (CPS), since it consists of many technologies such as computing, control and communications, and concerns both theoretical elements such as information theory in the context of control and practical issues such as software or hardware design. In this book, I focus only on the interplay of communications and control in CPSs, in particular how to design a communication system for the purpose of control, from the viewpoint of a communications engineer or scientist. Of course this does not mean that this book will benefit only the communications community. Experts on controls or computing, theorists or practical engineers may find interesting points in this book since it is a multidisciplinary subject.

This book consists of two parts. In the first part, which is composed of the first four chapters, the basics about CPS, communications, and control are briefly introduced. By reading through these four chapters, working knowledge can be acquired for readers unfamiliar with both communications and control. Then, the second part (the remaining four chapters), which is the main part of this book, explains various aspects of communication design for controlling CPSs, namely the capacity requirements, network topology design, networking, and physical layer design. These subjects are discussed specifically in the context of smart grids, using illustrated examples.

The writing of this book and many research results of mine incorporated in this book were supported by the National Science Foundation (NSF) under grants CNS-1239366, ECCS-1407679, CNS-1525418, and CNS-1543830. I sincerely thank the NSF for this generous support. I also sincerely thank my colleagues, Prof. Kevin Tomsovic and Prof. Seddik Djouadi, who collaborated with me on this NSF project and taught me the basics of power systems and automatic control. Last but not least, I would like to thank my family: without their full support, I could not have written this book.

Chapter 1

Introduction to cyber physical systems

Abstract

In this chapter, the big picture of the book is introduced. In particular, the key elements of cyber physical systems are explained. Then the key issues to be addressed, such as the communication requirements and communication system design, are outlined as an introduction to the contents of subsequent chapters.

Keywords

Cyber physical systems; Physical dynamics; Linear systems; Communications

1.1 Introduction

In the last decade, there have been numerous studies on cyber physical systems (CPSs), illustrated in Fig. 1.1. In CPSs, there exist physical, computing, and communication elements. To control the physical dynamics, sensors and controllers are connected via a communication network over which measurements on the system state can be conveyed. Many systems can be categorized as CPSs, such as smart grids, unmanned aerial vehicles, and robotic networks. The study on CPSs can provide a comprehensive and interdisciplinary framework for analyzing and designing these practical systems.

Fig. 1.1 Illustration of a CPS.

As a special example of CPS, smart grids may consist of a wide area monitoring system (WAMS), which measures the synchronized metrics such as voltage, currents, phase and frequency, and a set of controllers, e.g., the flexible AC transmission system [1]. In a microgrid, multiple distributed energy resources (DERs), such as wind turbines, are connected to a main power network; hence the power electronics interfaces of these DERs, which convert DC to AC power, should be controlled using the voltage measurements at the points of common coupling (PCCs; i.e., the connection point between the DER and the main power grid), which requires a communication link for the measurement feedback [2]. Networked control in smart grids can also be found in many other operations such as power consumption marketing (e.g., how to set a dynamic price for the electric power) and protective relays (which are triggered to isolate possible faulting devices when faults occur). In these CPSs, the communication infrastructure is of key importance since it plays a role analogous to that of the nervous system in the human body. When the communication system malfunctions, severe damage can be incurred to the dynamical system. Most existing studies on smart grids focus on isolated topics. Hence it is necessary to apply the framework of CPSs for holistic analysis and design.

1.2 Elements of a CPS

A typical CPS is illustrated in Fig. 1.1. It consists of the following components [3]:

• Physical dynamics: The P aspect of CPS is characterized by the physical dynamics (also called plant in the community of controls and systems), which could be the frequency fluctuations in power grids, or positions and velocities of robots, or vehicle densities along a highway. The physical dynamics change with time, and the evolution law is determined by the physical system itself and the control actions.

• Sensors: Sensors are used to sense the physical dynamics. The observations may be direct, i.e., the system state of physical dynamics, or indirect observations. There may be multiple sensors in a CPS. In smart grids, a sensor might be a synchronized phasor that samples the electric dynamics and computes the corresponding frequencies and phases.

• Controllers: A controller receives reports from sensors; it then computes and carries out the control action. Sometimes, the computation and actuation of the control action may be in different locations, thus needing communications. In this paper, we assume that the controller can carry out computation and actuation simultaneously. In smart grids, a frequency controller can compute the governor valve (which regulates the flow of steam pushing the mover) position and actuate it, based on the reports of frequencies in the power grid.

• Communication networks: If the sensors and controllers are not co-located, a communication network is needed for the sensors to send their reports to the controllers. The communication network can be either wired or wireless.

Dynamics of CPS

In a typical CPS such as power grids [4] and robotic networks [5], the physical dynamics can be described by a differential equation:

   (1.1)

where x is the system state, u is the control action, n is the physical law governing the evolution of system dynamics. If a discrete-time domain is considered, a difference equation can be used to describe the system in a similar manner:

   (1.2)

Such a differential equation description of dynamical systems has been used since the original use of Lagrangian mechanics and Hamiltonian mechanics, and has thus been studied intensively [6].

Note that there are other descriptions of physical dynamics, such as:

• Discrete-time systems such as Petri nets and stochastic timed automata, which are summarized in Ref. [7].

• Differential-algebraic equations, which consider both continuous-time and discrete events [8].

• Symbolic dynamical systems, which study the dynamics using graphs [9].

Since dynamical systems are a very broad area of study, it is prohibitively difficult to provide an exhaustive introduction to dynamical systems. A comprehensive introduction can be found in Ref. [10].

Linear Time-Invariant Systems

In many situations, e.g., when the system is time invariant and oscillates around the equilibrium point, the system dynamics can be simplified to linear time-invariant (LTI) dynamics either in the continuous-time domain:

   (1.3)

or in the discrete-time domain:

   (1.4)

, A, and B are time-invariant parameters of the system. The parameters of the continuous- and discrete-time systems are related by

   (1.5)

where Ts is the sampling period of the discrete-time system. Note that LTI systems are widely used in many practical scenarios due to their simplicity. Even if a system is nonlinear, it can still be approximated by a linear system near its operation point.

Observation Model

The controller(s) in a CPS needs observations on the system dynamics to estimate and control the system state. In the generic case, the observation is a function of the system state and random perturbations. For simplicity, we consider a linear observation model in this chapter:

   (1.6)

where y is the observation, w is random measurement noise, and C is the measurement matrix. Different sensors may observe different dimensions of y.

Feedback Control

The observation y . For simplicity we assume that a linear feedback control is used:

   (1.7)

where K is the feedback gain matrix. There are many approaches for the synthesis of K, e.g., the linear quadratic regulator (LQR) in which the cost function is a quadratic function of the system state and control actions. We will use LQR in our subsequent discussion due to its elegant framework and its capability to handle stochastic dynamics. The same principle can be applied in other control strategies such as model predictive control (MPC).

1.2.1 Communications for CPSs: Theoretical Studies

There have been several theoretical studies on the communications in CPSs from the viewpoint of networked control systems.

Communication requirements

A comprehensive introduction to the communication requirements in CPS, mainly in terms of communication channel capacities between the sensors and controllers, can be found in Ref. [11]. For the case of deterministic dynamical systems with an unknown initial state, it has been found that the requirement of the communication rate of sensors is dependent on the topological entropy [12] of the system dynamics [11]. The topological entropy is a metric measuring the uncertainty in the generation rate of a given dynamical system. Information theoretic arguments are also introduced to study the control through communication channels, such as the anytime capacity of a channel [13] and communication for stochastic linear control [14]. Another approach to analyze the communication requirements is to apply the second law of thermodynamics and use the reduction of entropy in the physical dynamics as a lower bound of the communication requirement [15].

Impact of communications on control

Given the communication network connecting the sensors and controllers in a CPS, the corresponding impacts of communication imperfections have been studied in the realm of networked control systems. The major impacts of communications include:

• Communication delay: Communication delay introduces memory to the system dynamics. Hence, extra variables need to be added to the system dynamics. When the communication delay is a constant, Ref. [16] provided the equivalent system dynamics, as well as the criterion for system stability. For the case of random delays, the corresponding system stability has been studied in Ref. [11].

• Packet drop: For random packet drops, the system can be considered as a Markovian jumping process whose equivalence can be found in Ref. [17]. Then, the stability conditions for Markovian jumping processes can be applied to analyze the system dynamics with random packet drops.

• Quantization: Quantization introduces errors into the measurement reports conveyed to the controllers. Hence the analysis can be based on small perturbations. Some related analysis, which analytically shows the impact of quantization on the control of a system, can be found in Ref. [18].

When more than one communication imperfection is present, the analysis will be more complicated, although still mathematically tractable [19]. For example, when there exist both significant packet drops and communication delay, the Markov jumping process can be applied to the equivalent system incorporating the communication delay, which makes the equivalent system complicated and possibly unsolvable.

Communication design

Seminal works have been carried out for communication system design in CPSs [20–22]. In Ref. [20], the MAC layer design is studied for distributed control systems, mainly based on numerical simulations. It is extended to the wireless communication in CPS within the framework of the linear quadratic Gaussian (LQG), in which the system dynamics are linear, the cost function is a quadratic one of the system state and control actions, and the random noise is assumed to be Gaussian [21]. In Ref. [22], the resource allocation in communication systems for CPS is formulated as an optimization problem with the awareness of the control system. In recent years, there have been more studies on this topic. For example, in Ref. [23], the IEEE 802.11 standard is applied and tuned (e.g., adjusting the network operation parameters such as the retransmission limit) for networked control systems. In Ref. [24], the routing problem (i.e., finding the path(s) from information source(s) such as sensors to information sink(s) such as controllers) is studied for the purpose of system state estimation. In Ref. [25], the communication topology is designed for distributed control, where the communication delay is omitted, which is reasonable when the delay is very small. In Ref. [26], various uncertainties, such as the uncertainties of delays and packet drops, in wireless networking endangering real-time estimation and control have been addressed.

1.2.2 Communications for CPS: Industrial Systems

Communication systems for industrial control systems are typical examples of those for CPS. Compared with commercial data communication systems such as cellular systems, the communication networks for industrial control systems have extra requirements such as predictable throughput, extremely low down time, operation in hostile environments, scalability and maintenance by other than communication specialists [27]. Below we introduce two types of popular communication systems for industrial controls. A comprehensive introduction can be found in Ref. [28].

• Foundation Fieldbus [29]: Fieldbus, a family of industrial control communication systems, has been widely used in industrial controls, which has been standardized as IEC61158. The Foundation Fieldbus is Type 1 of fieldbus. The largest deployment of the Foundation Fieldbus is China Nanhai and SECCO, each having about 15,000 fieldbus devices. It contains four layers, namely the Physical, Datalink, Application, and User layers. Different from many modern communication systems, the physical layer of fieldbus uses current modulation, in which each device senses the voltage drops at the terminating resistors. The Datalink layer manages the token passing for access control. Fieldbus is real-time and reliable, which is important for control systems. However, its data rate is low, around 30 kbps. Although it is very useful in many industrial controls with slow dynamics, it cannot be used in tasks having fast dynamics and requiring high data speed.

• Supervisory control and data acquisition (SCADA) [30]: The SCADA system is used to monitor and control the dynamics of industrial infrastructures such as power grids. It consists of a human-machine interface, supervisory computer system, remote terminal units (RTUs) connecting to sensors, and communication infrastructure. The communication network is either wired or wireless. For large area networks, SONET/SDH optical communications can be used. A disadvantage of SCADA is its slow speed, which may take minutes to collect data from the power grid [31]. Hence it cannot be used in fast data communications, such as data from phasor measurement units (PMUs) that measure the instantaneous phase and frequencies of different locations of the power grid, which are important sensors in power grids and will be explained later, and real-time control.

1.3 What is included and what is missing

In this section, we provide an outline of the book and also point out what has not been included.

1.3.1 Content

This book mainly consists of the following two parts:

• Basics of CPS: We give brief introductions to the basics of communications and controls in Chapters 2 and 3, thus providing working knowledge for the subsequent specific discussions. In Chapter 4, we introduce some typical CPSs, such as smart grids and aerial vehicles, which serve as working examples for the subsequent theoretical studies.

• Communication system design for CPS: Chapters 5-8 cover various topics on the design of communication systems in CPSs. It follows the order of system design. We begin with the theoretical analysis on the communication capacity requirements for the purpose of control in Chapter 5. Once the communication capacity requirement is well understood, we then consider how to design the communication network topology in Chapter 6, based on the requirement of throughput. Given a fixed communication network topology, we study the networking protocol in CPS in Chapter 7, ranging from the MAC layer to the network layer. Finally we study the physical layer design for communications in CPS, such as modulation, source coding, and channel coding. The key feature of these studies is the incorporation of the characteristics of control and physical dynamics in the design procedure of communications, thus resulting in a physical dynamics aware communication system.

1.3.2 Missing Topics

This book is far from an exhaustive introduction to CPSs. The following aspects of CPSs are not included in this book, due to the scope of the authors’ research and limitations of space:

• Computing in CPS: It is well recognized that computing, control, and communications are interleaved and integrated in CPS. The computing for control, and communication for computing, and the overall interactions of these three components are important for many typical CPSs. There have been numerous studies on communications for computing (e.g., the theory of communication complexity), and also some studies on the computing aspects of control and decision making (e.g., the computational complexity of Markov decision processes). However, there have been no significant studies on the corresponding interactions in the context of CPS.

• Sensor networks: This is an important topic since a CPS needs to use sensors to measure physical dynamics. There are many research issues in sensor networks, e.g., the power consumption schedule due to the limited battery power of many sensors, various networking issues (such as scheduling and routing) for communications, security issues (such as data privacy), and in-network computation, which saves communication resources. Since there have been many monographs on sensor networks, we do not provide a systematic introduction of this topic in this book.

• Security issues: Information security such as data privacy and message authentication. There have been some studies on this aspect of CPSs; e.g., data injection attack in smart grids, Global Positioning System (GPS) spoofing attacks in PMUs of power systems, and resilient control for generic CPSs. Many existing techniques in the area of cryptography, such as oblivious transfer and differential privacy, can be applied; however, these generic approaches have not been thoroughly studied in the context of CPSs.

References

[1] Phadke A.G., Thorp J.S. Synchronized Phasor Measurements and Their Applications. New York: Springer; 2008.

[2] Li H., Li F., Xu Y., Rizy D.T., Kueck J.D. Adaptive voltage control with distributed energy resources: algorithm, theoretical analysis, simulation, and field test verification. IEEE Trans. Power Syst. 2010;25(3).

[3] Lee E.A. Cyber physical systems: design challenges. In: Proceedings of the IEEE International Symposium on Object Oriented Real-Time Distributed Computing (ISORC). 2008.

[4] Kundur P. Power System Stability and Control. McGraw-Hill; 1994.

[5] Bullo F., Cortés J., Martínez S. Distributed Control of Robotic Networks. Princeton University Press; 2009.

[6] Arnold V.I., Weinstein A. Mathematical Methods of Classical Mechanics. New York: Springer; 2007.

[7] Cassandras C.G., Lafortune S. Introduction to Discrete Event Systems. Cambridge, UK: Cambridge University Press; 2009.

[8] Kunkel P., Mehrmann V. Differential-Algebraic Equations: Analysis and Numerical Solutions. European Mathematical Society; 2006.

[9] Lind D., Marcus B. An Introduction to Symbolic Dynamics and Coding. Cambridge, UK: Cambridge University Press; 1995.

[10] Katok A., Hasselblatt B. Introduction to Modern Theory of Dynamical Systems. Cambridge, UK: Cambridge University Press; 1995.

[11] Matveev A.S., Savkin A.V. Estimation and Control Over Communication Networks. Birkhäuser; 2008.

[12] Downarowicz T. Entropy in Dynamical Systems. Cambridge, UK: Cambridge University Press; 2011.

[13] Sahai A., Mitter S. The necessity and sufficiency of anytime capacity for control over a noisy communication channel. IEEE Trans. Inform. Theory. 2006;52(8):3369–3395.

[14] Tatikonda S., Sahai A., Mitter S. Stochastic linear control over a communication channel. IEEE Trans. Inform. Theory. 2004;49(9):1549–1561.

[15] Li H. Entropy reduction via communications in cyber physical systems: how to feed Maxwell’s Demon? In: Proceedings of the IEEE International Symposium of Information Theory. 2015.

[16] Branicky M.S., Phillips S.M., Zhang W. Stability of networked control systems: explicit analysis of delay. In: Proceedings of the American Control Conference (ACC). 2000.

[17] Seiler P., Sengupta R. Analysis of communication losses in vehicle control problems. In: Proceedings of the American Control Conference (ACC). 2001.

[18] Montestruque L.A., Antsakilis P.J. Quantization in model based networked control systems. In: Proceedings of the 16th IFAC World Congress. 2005.

[19] Hespanha J.P., Naghshtabrizi P., Xu Y. A survey of recent results in NCS. Proc. IEEE. 2007;95:138–162.

[20] Liu X., Goldsmith A. Wireless medium access control in distributed control systems. In: Proceedings of the Annual Allerton Conference on Communications, Control and Computing, Monticello, IL; 2003.

[21] Liu X., Goldsmith A. Wireless network design for distributed control. In: Proceedings of the IEEE Conference on Decision and Control (CDC), Paradise Island, Bahamas; 2004.

[22] Xiao L., Johansson M., Hindi H., Boyd S., Goldsmith A. Joint optimization of communication rates and linear systems. IEEE Trans. Automat. Control. 2003;48:148–153.

[23] Bai J., Eyisi E.P., Xue Y., Koutsoukos X.D. Dynamic tuning retransmission limit of IEEE 802.11 MAC protocol for networked control systems. In: Proceedings of the First International Workshop on Cyber-Physical Networking Systems (CPNS). 2011.

[24] Gupta V. On an estimation oriented routing protocol. In: Proceedings of the American Control Conference (ACC), Baltimore; 2010:580–585.

[25] Langbort C., Gupta V. Minimal interconnection topology in distributed control design. In: Proceedings of the American Control Conference (ACC), Minneapolis; 2006.

[26] Liu X., Zhang H., Xiang Q., Che X., Ju X. Taming uncertainties in real-time routing for wireless networked sensing and control. In: Proceedings of the 13th ACM International Symposium on Mobiel Ad Hoc Networking and Computing (MobiHoc), Head Island, USA; 2012.

[27] Thompson L.M. Industrial Data Communications. fourth ed. Instrumentation Systems; 2002.

[28] Galloway B., Hancke G. Introduction to industrial control networks. IEEE Commun. Surv. Tut. 2012;99:1–21.

[29] National Instruments, Fieldbus: Foundation Fieldbus Overview, 2003.

[30] Boyer S.A. SCADA: Supervisory Control and Data Acquisition. fourth ed. The Instrumentation, Systems and Automation Society; 2009.

[31] Mynam M.V., Harikrishna A., Singh V. Synchrophasors Redefining SCADA Systems. Schweitzer Engineering Laboratories, Inc., 2013.

☆ "To view the full reference list for the book, click here"

Chapter 2

Basics of communications

Abstract

This chapter provides a brief introduction to the basics of communications, for the benefit of readers