Networked Triggered Switched Systems Control Under DoS Attacks

There are two major problems in the design of networked control: 1)the limitation of network communication resources;2)the security of network data transmission.For problem1),the event-triggering mechanism that realizes efficient configuration of network resources by means of on-demand transmission is an effective solution.For problem 2),it is an important research idea to design the attack detection and compensation mechanism to actively defense the attack,or to improve the existing control framework to obtain the maximum tolerance.However,the current research on the event-triggered control of switched systems under network attacks is still preliminary in the field of control and remains to be explored.Based on this,this thesis studies the control problem of triggered networked switched systems under the typical and important type of attack(DoS attack).First,this thesis investigates the research status of networked control switched systems.Then,aiming at the complex interleaving of attack,triggering and switching time in the triggered switched system under DoS attack,the coupling relationship is analyzed in detail,and an resilient event-triggering mechanism is introduced to actively reduce the impact of DoS attack.The sufficient conditions for ensuring the stability of the system are deduced by the average dwell time technique,and the LMI toolbox is used to jointly solve the controller gain and event-triggering parameters.Then,aiming at the problem of switching asynchronous caused by network delay and DoS attack,active confrontation and passive tolerance mechanisms are designed to combat the mismatch of system modes.In this design,a combined new switching law is constructed to eliminate the asynchronous phenomenon,an adaptive elastic event-triggering mechanism is proposed to improve the ability to configure network resources,and sufficient conditions to ensure system stability and the maximum tolerance of DoS attacks are given.In passive design,the timing relationship between asynchronous,expected synchronization and unexpected synchronization interval is analyzed in detail,and the co-design method of switching law,controller gains and event-triggering parameters under periodic DoS attack is given.Furthermore,aiming at the optimal control problem of the triggered switching system under DoS attack,the adaptive dynamic optimization and adaptive dynamic programming control methods are studied.For the existing problem of system data damage under attack,a predictive compensation based on neural network is designed.A mechanism is used to ensure the training accuracy of the control network,and the convergence and optimality of the proposed iterative algorithm are proved.In addition,based on the Matlab/Simulink platform,this thesis uses the numerical simulation method to verify the correctness and effectiveness of each proposed theoretical method.