Research On The Safety Consistency Control Of A Class Of Multi-agent Systems Under Network Attack

Aiming at networked multi-agent systems under cyber-attacks,this paper studies the secure consensus control problem.For several common cyber-attacks,the unified model framework has been established,which can systematically describe the effects of single or mixed attacks.In addition,in order to reduce the communication burden among agents and save energy,the dynamic event-triggered algorithm is introduced.The consensus controllers are designed to ensure the desired consensus performance of multiagent systems under cyber-at tacks.With the help of convex optimization technology and random analysis methods,sufficient conditions for the existence of the required controller are given,and the controllers gains are given by solving a set of recursive linear matrix inequalities.Specifically,the research contents of this paper include the following aspects:First,the research background and significance of this paper are introduced,and the research status and progress of multi-agent systems are summarized.Subsequently,the main content and structure arrangement are described.Second,chapter 2 studies the consensus control problem of a kind of time-varying stochastic multi-agent systems subject to data injection attack.The probability-guaranteed concept is introduced,and the principle of probability-guaranteed consensus is proposed.With the help of convex optimization technology and random analysis methods,the sufficient conditions for the existence of the expected consensus protocol are given,so that the states of each agent are constrained to a desired ellipsoid with a certain probability.Subsequently,two optimization algorithms are proposed to characterize the local optimal consensus performance of the multi-agent system.A simulation example verifies the effectiveness of the proposed algorithm.Third,chapter 3 studies the consensus control problem of a class of time-varying stochastic multi-agent systems subject to mixed attacks and sensor saturation.The additive random noise and multiplicative random noise are introduced to describe the random interference experienced by the agent.A mixed attack model is established,which includes two common types of attacks,namely data injection attacks and replay attacks.The consensus criterion of probability-guaranteed is proposed.Using convex optimization technology and stochastic analysis method,sufficient conditions for the existence of the expected consensus controller are obtained.Subsequently,two optimization algorithms for finding local consensus performance are proposed.The simulation example verifies the effectiveness of the proposed safety control algorithm.Then,chapter 4 investigates the dynamic event-based H∞ consensus control problem of a class of multi-agent systems under multiple attacks.A multiple attacks model is established,including scaling attacks,denial of service attacks,and replay attacks.In order to reduce the communication burden,the dynamic event-triggered mechanism is proposed.The sufficient conditions for the existence of the expected controller are given.Numerical examples show the effectiveness and correctness of the safety control scheme.Finally,the main work of this research is summarized,and the future research direction is prospected.