On Switching Control Of Cyber-Physical Systems Under DoS Attacks

With the advent of 5G era,human society is gradually stepping into a fully connected,intelligent,digital era.Meanwhile,by upgrading traditional industrial control systems with the latest information and communications technology,a new generation of industrial control systems—cyber-physical systems(CPS)represented by smart grid,intelligent transportation,smart factory,etc.,has emerged.In CPS,different function modules need to use proprietary or public networks to exchange information,which inevitably exposes the CPS to various network attacks.Since many applications of CPS are safety-critical,any successful attacks on CPS may result in catastrophic consequences.Therefore,it is significant to investigate the secure control problem of CPS under network attacks.From the perspective of switching system theory,some problems about the modeling,state estimation and secure control of CPS under DoS attacks are investigated in this thesis.The specific research contents of this thesis are summarized as follows:1.The secure control problem of CPS under energy-constrained DoS attacks is investigated.First,the dynamics of the CPS under DoS attacks with different durations are modeled by the switching system model.Thus,the secure control problem of the CPS under DoS attacks is transformed into the stability problem of switching system.Finally,by resorting to multiple Lyapunov function method,several sufficient conditions are derived such that the switching system is exponentially stable under arbitrary switching laws.Meanwhile,the corresponding controller design algorithm with input constraint is derived such that the influence of DoS attacks can be compensated swiftly.2.For the secure control problem of CPS under energy-constrained DoS attacks,a redundant control method is proposed.In the absence of DoS attacks,the current control signals along with some redundant control signals are encapsulated into one packet and sent to the actuators simultaneously such that the actuators always have appropriate control signals to be applied no matter what the DoS attacks occur or not.Finally,by constructing a nested switching system model,some sufficient stability conditions are derived.Meanwhile,the controller design algorithm is also provided.The proposed redundant control method overcomes the blindness of the zero-input or hold-input strategies and achieves the resilient control under DoS attacks.3.The secure control problem of CPS with partial unmeasurable states under energy-constrainted DoS attacks is investigated.First,a switching Luenberger observer with varying gains is proposed to obtain the intermittent state estimations of CPS under DoS attacks.Then,a redundant control law is propsed based on the obtained state estimations.Further,by introducing an augmented variable,the dynamics of system states and observer states are modeled by a switching system model simultaneously.Finally,by using multiple Lyapunov function method and cone complementarity linearization algorithm,the exponential stability conditions and the design algorithm of the observer and controller are derived.Thus,the observer-based controller design of CPS under energy-constrained DoS is achieved.4.In the end,some simulation analyses and experiment are conducted to demonstrate the effectiveness of the proposed method.Meanwhile,the thesis is concluded and some directions of future works are discussed.