Skip to main content
main-content

Über dieses Buch

This book introduces a cross-layer design to achieve security and resilience for CPSs (Cyber-Physical Systems). The authors interconnect various technical tools and methods to capture the different properties between cyber and physical layers. Part II of this book bridges the gap between cryptography and control-theoretic tools. It develops a bespoke crypto-control framework to address security and resiliency in control and estimation problems where the outsourcing of computations is possible. Part III of this book bridges the gap between game theory and control theory and develops interdependent impact-aware security defense strategies and cyber-aware resilient control strategies.

With the rapid development of smart cities, there is a growing need to integrate the physical systems, ranging from large-scale infrastructures to small embedded systems, with networked communications. The integration of the physical and cyber systems forms Cyber-Physical Systems (CPSs), enabling the use of digital information and control technologies to improve the monitoring, operation, and planning of the systems. Despite these advantages, they are vulnerable to cyber-physical attacks, which aim to damage the physical layer through the cyber network.

This book also uses case studies from autonomous systems, communication-based train control systems, cyber manufacturing, and robotic systems to illustrate the proposed methodologies. These case studies aim to motivate readers to adopt a cross-layer system perspective toward security and resilience issues of large and complex systems and develop domain-specific solutions to address CPS challenges.

A comprehensive suite of solutions to a broad range of technical challenges in secure and resilient control systems are described in this book (many of the findings in this book are useful to anyone working in cybersecurity). Researchers, professors, and advanced-level students working in computer science and engineering will find this book useful as a reference or secondary text. Industry professionals and military workers interested in cybersecurity will also want to purchase this book.

Inhaltsverzeichnis

Frontmatter

Motivation and Framework

Frontmatter

Chapter 1. Introduction

Abstract
Cyber-physical systems (CPSs) are systems that integrate sensing, computation, control, and networking into physical systems. They have transformed the way we interact with engineered systems and herald many emerging technologies and applications, including smart grids, connected and autonomous vehicles, cyber manufacturing, and intelligent healthcare systems. One essential challenge of designing CPSs is the interdependencies between the cyber and the physical layers of the system. In particular, the strong coupling between the two layers provides adversaries new opportunities to compromise the physical assets through the cyber system. Traditional methods, such as cryptography and intrusion detection systems, are not sufficient to safeguard the CPSs. Instead, it is vital to develop a cross-layer approach that enables interdependent impact-aware security defense strategies and cyber-aware resilient control strategies. To this end, this book presents two methodologies for the design. The first one focuses on cryptography and control theory, which are used to address control and estimation problems where the outsourcing of computations is possible. The second one focuses on the interconnection between game theory and control theory to resolve design problems with explicit attack models. This chapter gives an overview of the book and discusses literature related to the proposed cross-layer approach.
Quanyan Zhu, Zhiheng Xu

Chapter 2. Cross-Layer Framework for CPSs

Abstract
CPSs are composed of two interdependent layers. One is the cyber layer, and the other is the physical layer. The physical system becomes vulnerable as there are growing connectivity and integration between the cyber and the physical systems. Traditional IT solutions are not sufficient to protect the CPS from increasingly sophisticated cyber attacks. This chapter introduces the conceptual frameworks for cross-layer CPS design to improve the security at the cyber layer and the resiliency at the physical layer in a holistic manner. Based on the general conceptual framework, this chapter introduces two methodologies to design secure and resilient CPS. The first one builds an integrated framework bridging cryptography for data privacy and integrity with control theory for stabilizing the physical systems. The cross-layer design enables the cryptographic solution to be aware of its impact on the physical systems and the control-theoretic design to be aware of the consequences of cryptographic solutions. The second one provides a unified design paradigm using game theory. Game theory provides a rich class of models that can be used to capture multiple types of interactions. It can be used to model the adversarial behaviors and their interactions with the cyber defense, as well as the optimal robust control design under worst-case disturbances. The game-theoretic unification of the models at the cyber and the physical layers naturally leads to a holistic cross-layer design framework. This chapter introduces the concepts of games-in-games, or meta-game to describe the system-of-systems modeling of CPSs. We extend the cross-layer framework from a complete observation paradigm to a partially observable one. We will use multiple CPS application domains to illustrate the two cross-layer design methodologies.
Quanyan Zhu, Zhiheng Xu

Secure Outsourcing Computations of CPS

Frontmatter

Chapter 3. New Architecture: Cloud-Enabled CPS

Abstract
Large-scale CPSs gather an increasing amount of data from networked and distributed sensors. They require a significant amount of computational resources to process the data, extract the information, and control the plants to achieve system objectives. As a result, resource-constrained control systems are not able to attain these objectives with conventional system architectures. For example, in the case of nano and microrobots, the size of the physical part of the robots is often limited. It constrains the hardware’s resources, making it impractical for the robots to handle complex sensing information, such as video and image information. The recent development of Cloud Computing Technologies (CCTs) makes it possible to resolve the resource-constrained issues of CPSs. CCTs provide massive computing resources and storage resources to CPSs and extensively enhance their performance by migrating massive local computations to a cloud. This chapter briefly introduces the advantages and basic framework of the CE-CPS.
Quanyan Zhu, Zhiheng Xu

Chapter 4. Secure and Resilient Design of Could-Enabled CPS

Abstract
Cloud computing enables resource-constrained CPSs to outsource heavy computations to a cloud server with massive computational resources. However, Cloud-Enabled CPSs introduce new challenges arising from the trustworthiness of the cloud and the cyber-physical connections between the control system and the cloud. To address these concerns, this chapter presents a secure and resilient mechanism, which employs customized cryptographic tools to encrypt the data of a control problem and develops verification methods to guarantee the integrity of the computational results from the cloud. In addition, our design enables a Switching Mode Mechanism (SMM) to provide resiliency to the CPSs when the system successively fails to receive correct control inputs from the cloud. We demonstrate that the mechanism can achieve data confidentiality and integrity, guarantee the stability, and enhance the resiliency. Finally, an Unmanned Aerial Vehicle (UAV) example is used to corroborate these properties.
Quanyan Zhu, Zhiheng Xu

Chapter 5. Secure Data Assimilation of Cloud Sensor Networks

Abstract
Cloud computing technologies (CCTs) enable a large-scale sensor network (LSN) to outsource the computations of data assimilation to improve its performance. However, the cyber-physical nature of cloud-enabled LSNs (CE-LSNs) introduces new challenges. Outsourcing the computations to an untrusted cloud may expose the privacy of the sensing data. To address the security issues, this chapter proposes a security mechanism to achieve data confidentiality in the outsourcing process. We develop our mechanism by combining a conventional homomorphic encryption and a customized encryption scheme. We present theorems to characterize the correctness of the encryption and investigate the estimation performance and the security of the proposed method. We also analyze the impacts of the quantization errors on the estimation performance. Finally, we present numerical experiments to consolidate our analytical results.
Quanyan Zhu, Zhiheng Xu

Game-Theoretic Approach for CPS

Frontmatter

Chapter 6. Review of Game Theory

Abstract
Theoretic models, such as game-theoretic model, Markov decision processes, linear and non-linear dynamic models, etc., at the system level, play a significant role in Cyber-Physical Systems (CPSs). These models stand as a scientific basis for high-level, security-related decision making. Game theory provides mathematical tools and models for investigating multi-person strategic decision making, where the players or decision-makers compete for limited and shared resources (Manshaei et al., ACM Comput Surv 45(3):25, 2013). Considering the security issues in CPSs, we can view the defenders and attackers as two types of players, who fight for specific resources. Therefore, we use game models to describe the interactions between them. The analytical results based on the game models can assist us in finding the optimal defense strategy. In this chapter, we briefly review specific game theory models, including zero-sum games, Stackelberg games, and FlipIt game. These models are closely related to our applications.
Quanyan Zhu, Zhiheng Xu

Chapter 7. A Game-Theoretic Approach to Secure Control of 3D Printers

Abstract
Due to the high costs of 3D-printing infrastructure, outsourcing the production to third parties specializing in the 3D-printing process becomes necessary. The integration of a 3D-printing system with networked communications constitutes a cyber-physical system, bringing new security challenges. Adversaries can explore the vulnerabilities of networks to damage the physical parts of the system. In this chapter, we explore the vulnerabilities of 3D-printing systems and design a cross-layer approach for the system. At the physical layer, we use a Markov jump system to model the system and develop a robust control policy to deal with uncertainties. At the cyber-layer, we apply FlipIt game to model the contention between the defender and attacker for the control of the 3D-printing system. To connect these two layers, we develop a Stackelberg framework to capture the interactions between the cyber-layer attacker and defender game and the physical-layer controller and disturbance game, and define a new equilibrium concept that captures interdependence of the zero-sum and FlipIt games.
Quanyan Zhu, Zhiheng Xu

Chapter 8. A Game Framework to Secure Control of CBTC Systems

Abstract
To meet the growing railway-transportation demand, researchers have designed a new train control system, communication-based train control (CBTC) system, to maximize the ability of train lines by reducing the headway of each train. However, wireless communications expose the CBTC system to new security threats. Due to the cyber-physical nature of the CBTC system, a jamming attack can damage the physical part of the train system by disrupting the communications. To address this issue, this chapter develops a secure framework to mitigate the impact of the jamming attack based on a security criterion. At the cyber layer, we apply a multi-channel model to enhance the reliability of the communications and develop a zero-sum stochastic game to capture the interactions between the transmitter and jammer. We present analytical results and use dynamic programming to find the equilibrium of the stochastic game. We investigate a special case study to obtain insight results, showing a strong inter-dependency between the physical and cyber layers. Finally, the experimental results are provided to evaluate the performance of the proposed secure mechanism.
Quanyan Zhu, Zhiheng Xu

Chapter 9. Secure Estimation of CPS with a Digital Twin

Abstract
Cyber-Physical Systems (CPSs) play an increasingly significant role in many critical applications. These valuable applications attract various sophisticated attacks. This chapter considers a stealthy estimation attack, which aims to modify the state estimation of the CPSs. The intelligent attackers can learn defense strategies and use clandestine attack strategies to avoid detection. To address the issue, we design a Chi-square detector in a Digital Twin (DT), which is an online digital model of the physical system. We use a Signaling Game with Evidence (SGE) to find the optimal attack and defense strategies. Our analytical results show that the proposed defense strategies can mitigate the impact of the attack on the physical estimation and guarantee the stability of the CPSs. Finally, we use an illustrative application to evaluate the performance of the proposed framework.
Quanyan Zhu, Zhiheng Xu

Chapter 10. Introduction to Partially Observed MDPs

Abstract
In our cross-layer design, we use different models to capture the properties of different layers. As stated in Chap. 8, we can use an MDP model to capture the dynamical movements of the cyber layer. However, in the scenario, we assume that the defender can observe the cyber state at each cyber time instant. In real applications, it is challenging to obtain the full information of the cyber state directly. Hence, the MDP cannot capture the incomplete knowledge of the cyber states. In this chapter, we will introduce a Partially Observed Markov Decision Process (POMDP) to capture the uncertainty of the cyber state. In a POMDP, instead of observing the states, we have an observation, whose distribution depends on the state. Therefore, we use this information to build a Hidden Markov Model (HMM) filter, which can construct a belief of the states. Based on the belief, we aim to find an optimal policy to minimize an expected cost.
Quanyan Zhu, Zhiheng Xu

Chapter 11. Secure and Resilient Control of ROSs

Abstract
A Robot Operating System (ROS) plays a significant role in organizing industrial robots for manufacturing. With an increasing number of robots, operators integrate a ROS with networked communications to share the data. This cyber-physical nature exposes the ROS to cyber attacks. This chapter proposes an integrated cyber-physical solution to secure control of ROS agents using impact-aware lightweight cryptography and cyber-aware control design. We use the cyber states to capture the well-being of the cyber system under the lightweight encryption scheme. To model the incomplete information of the defender on the cyber state, we use a Partially Observed Markov Decision Process (POMDP) to design an impact-aware defense mechanism that makes the ROS resilient to the attacks and mitigates the cyber-physical risks. Finally, we show that the proposed scheme provides a high-confidence control of delay-sensitive ROS in the numerical experiments. The results also show a strong inter-dependency between physical parameters and cyber performance.
Quanyan Zhu, Zhiheng Xu

Discussion of the Future Work

Frontmatter

Chapter 12. Future Work in Security Design of CPSs

Abstract
The increasing reliance on computer and communication technologies exposes control systems to cybersecurity threats. The physical systems can now be attacked through cyberspace. Emerging sophisticated attacks can exploit zero-day vulnerabilities, persist in the system for long periods of time, and advance stealthily to achieve their attack goals. Protection and prevention against such attacks are not always possible, and a paradigm shift to emphasize the resilience of a control system is the overarching objective for safeguarding control systems to protect the nation’s critical infrastructures. This book has developed a set of powerful and versatile tools to establish scientific co-design principles for integrating cyber and physical systems in a holistic manner. The methodologies in this project provide a holistic and interdisciplinary approach to the secure and resilient CPS design, bridging the gap between multiple disciplines such as computer science, control, communications, and economics. In this chapter, we present several problems that could be addressed by applying or extending the proposed methodologies, and discuss future research directions of the secure and resilient CPSs. We first present two attack models for CPSs, including the man-in-the-middle (MITM) attack and the compromised-key attacks. The cross-layer co-design framework can potentially address these two attack models. Most of the existing works have focused on data confidentiality and integrity. We argue that data availability plays a critical role in time-critical CPSs. There is a need to use the cross-layer design framework to design safe-mode mechanisms and defense strategies to improve data availability.
Quanyan Zhu, Zhiheng Xu

Backmatter

Weitere Informationen

Premium Partner

    Bildnachweise