Security of Cyber-Physical Systems: State Estimation and Control

✍ Scribed by Chengwei Wu, Weiran Yao, Guanghui Sun, Ligang Wu

• Publisher: Springer
• Year: 2021
• Tongue: English
• Leaves: 293
• Series: Studies in Systems, Decision and Control 396
• Edition: 1
• Category: Library

✦ Synopsis

This book analyzes the secure problems of cyber-physical systems from both the adversary and defender sides. Targeting the challenging security problems of cyber-physical systems under malicious attacks, this book presents some recent novel secure state estimation and control algorithms, in which moving target defense scheme, zero-sum game-theoretical approach, reinforcement learning, neural networks, and intelligent control are adopted. Readers will find not only the valuable secure state estimation and control schemes combined with the approaches aforementioned, but also some vital conclusions for securing cyber-physical systems, for example, the critical value of allowed attack probability, the maximum number of sensors to be attacked, etc. The book also provides practical applications, example of which are unmanned aerial vehicles, interruptible power system, and robot arm to validate the proposed secure algorithms. Given its scope, it offers a valuable resource for undergraduate and graduate students, academics, scientists, and engineers who are working in this field.

✦ Table of Contents

Preface
Acknowledgements
Contents
Notations and Acronyms
List of Figures
List of Tables
1 Introduction
1.1 Current Research on Securing CPSs
1.1.1 Advances in Detecting Attacks
1.1.2 Advances in Securing CPSs Under DoS Attacks
1.1.3 Advances in Securing CPSs Under Deception Attacks
1.2 Publication Contribution
1.3 Publication Outline
Part I Secure Estimation and Control for CPSs Under DoS Attacks
2 Optimal DoS Attack Scheduling for CPSs
2.1 Introduction
2.2 System Formulation and Preliminaries
2.2.1 Physical Process Description
2.2.2 SINR-Based Communication Model
2.3 Remote Estimator Design
2.4 Attack Sequence Design and Energy Allocation
2.5 Simulation Results
2.6 Conclusion
3 Active Defense Control of CPSs via Sliding Mode
3.1 Introduction
3.2 System Formulation and Preliminaries
3.2.1 Physical Process Description
3.2.2 DoS Attack
3.3 Resilient Sliding Mode Control Design
3.3.1 Estimator Design
3.3.2 Sliding Mode Approach
3.3.3 Stability Analysis
3.3.4 Resilient Sliding Mode Controller Design
3.4 Active Defense-Based Resilient Control
3.4.1 Optimal Defense Strategy
3.4.2 Defense Scheme-Based Estimator
3.4.3 Sliding Dynamics Analysis
3.4.4 Resilient Sliding Mode Control Design with Defense Strategy
3.5 Conclusion
4 Learning Tracking Control for CPSs
4.1 Introduction
4.2 System Formulation and Preliminaries
4.2.1 Physical Process and Reference Model Descriptions
4.2.2 SINR-Based Communication Model
4.2.3 Control Objective
4.3 Optimal Tracking Controller Design and Stability Analysis
4.3.1 Analysis of the Value Function
4.3.2 Optimal Tracking Controller Design
4.3.3 Analysis of the Solution to the Riccati Equation
4.4 Q-Learning Optimal Tracking Scheme Design
4.4.1 Convergence Analysis of Algorithm
4.5 Off-Policy Learning Control Algorithm
4.5.1 Model-Based Off-Policy Learning Control Scheme
4.5.2 Model-Free Off-Policy Learning Control Scheme
4.6 Conclusion
5 Intelligent Control for Nonlinear Networked Control Systems
5.1 Introduction
5.2 Problem Formulation
5.2.1 System Description
5.2.2 Fuzzy Logic Systems
5.3 Adaptive Control Design
5.4 Simulation Results
5.5 Conclusion
6 Reliable Filtering of Sensor Networks
6.1 Introduction
6.2 Problem Formulation
6.2.1 IT-2T-S Fuzzy Model
6.2.2 Failure Model
6.2.3 IT2 Fuzzy Reliable Distributed Filter Design
6.2.4 Filtering Error System
6.3 Main Results
6.4 Simulation Results
6.5 Conclusion
Part II Secure Estimation and Control for CPSs Under False Data Injection Attacks
7 Secure Estimation for CPSs via Sliding Mode
7.1 Introduction
7.2 Problem Formulation
7.2.1 CPSs Model
7.2.2 CPSs Model in Descriptor Form
7.2.3 Minimization Optimal Problem
7.2.4 Sliding Mode Observer
7.3 Secure Estimation Algorithm Design
7.3.1 Projection Operator
7.3.2 Sliding Mode Observer Secure Estimation Algorithm
7.4 Convergence Analysis of Algorithm 1
7.4.1 Analysis for Time Update
7.4.2 Analysis for Projection Operator
7.4.3 Analysis for Estimation Update
7.4.4 Analysis for Terminated Conditions of Inner Loop
7.4.5 Analysis of Sliding Mode Error Rejection Term Design
7.5 Simulation Results
7.6 Conclusion
8 Zero-Sum Game Based Optimal Secure Control
8.1 Introduction
8.2 Problem Formulation
8.2.1 Physical Process Description
8.2.2 Actuator Attack
8.3 Main Results
8.3.1 Optimal Estimator Design
8.3.2 Zero-Sum Game Based Optimal Policies Design
8.3.3 Convergence of the Scheme
8.4 Extended Work
8.5 Simulation Results
8.6 Conclusion
9 Proactive Secure Control for CPSs
9.1 Introduction
9.2 System Formulation and Preliminaries
9.2.1 Physical Process: A Switching System Representation
9.2.2 The Closed-Loop System Under Attacks
9.3 Moving Target Defense Control Scheme
9.3.1 Controller and Moving Target Defense Design
9.3.2 Stability Analysis
9.4 Attack Detection and Isolation
9.4.1 Attack Detection Observer Design
9.4.2 Attack Isolation Scheme
9.5 Moving Target Defense Control Design
9.5.1 Reinforcement Learning Based Reactive Control Scheme
9.5.2 Moving Target Defense Design
9.6 Simulation Results
9.7 Conclusion
10 Fault-Tolerant Tracking Control for Nonstrict-Feedback Systems
10.1 Introduction
10.2 System Description and Preliminaries
10.2.1 Nonstrict-Feedback System
10.2.2 Fault Model
10.2.3 Control Objective
10.2.4 Neural Networks
10.2.5 Observer Design
10.3 Observer-Based Adaptive Fault-Tolerant Control Design
10.4 Simulation Results
10.5 Conclusion
11 Deep Reinforcement Learning Control Approach to Mitigating Attacks
11.1 Introduction
11.2 System Formulation and Preliminaries
11.2.1 Physical System Description
11.2.2 CPSs Under Cyber Attacks
11.3 Deep Reinforcement Learning Based Secure Controller Design and Implementation
11.3.1 Markov Decision Process
11.3.2 Reinforcement Learning Algorithm
11.3.3 Deep Neural Networks Approximation
11.3.4 Implementation of Lyapupnov-Based Soft Actor-Critic Deep Reinforcement Learning Control Algorithm
11.4 Convergence and Stability Analysis
11.4.1 Algorithm Convergence Analysis
11.4.2 Data-Based Stability Analysis
11.5 Simulation Results
11.6 Conclusion
12 Conclusion and Further Work
12.1 Conclusion
12.2 Further Work
Appendix References