Robust H_∞ Control Of Several Complex Systems Under Bandwidth Constraints And Network Attacks

Networked control systems(NCSs)are applied in various fields based on their characteristics.The increasing scale and complexity of networked systems have led to many prominent problems.First,with the continuous increase of data transmission in the networked control systems,the inability of limited network bandwidth to support the transmission of a large amount of data will cause data redundancy and network congestion,and lead to the deterioration of system performance.Second,the process of network data transmission may undergo network attacks,which will invade and destroy the control system and cause huge losses to the practical industrial production and other aspects.Third,the uncertainty,time-varying,nonlinear and other characteristics of system parameters increase the difficulty of building relevant mathematical models.This paper conducts a study on the robust control problems for different networked control systems from the perspective of network bandwidth limitation by taking network security factors into consideration.The main contents of this paper are as follows:Firstly,the problem of H_∞ dynamic output feedback control for discrete networked Markov jump systems with uncertainties was studied.A two-terminal trigger mechanism was designed to be placed between the sensor-controller and the controller-actuator data transmission channels to save network resources and network bandwidth resources.At the same time,the design method of the dynamic output feedback controller was given,and sufficient conditions for the closed-loop control system to be stochastic stable and meet the H_∞performance index were derived.The parameters of the dynamic output feedback controller were obtained by the congruent transformation method,and the effectiveness of the scheme was verified by numerical simulation.Then,under the hybrid-trigger mechanism consisting of event-trigger and time-trigger,the H_∞control problem was studied for discrete stochastic systems and discretely distributed delay Markov jump systems respectively.Controllers were designed for the Markov jump system respectively under the two data transmission modes of hybrid-trigger.Accordingly,a less conservative double-gain control scheme was developed,and a closed-loop control system with double random characteristics was constructed.In addition,LMI technology was applied to transform the controller design problem into a H_∞performance optimization problem.The simulation result verified the effectiveness of the controller design and the comparison results of different trigger mechanisms showed the advantages of the design scheme.Thirdly,considering the factors of network security,to further save network resources,the time-trigger mechanism,and adaptive-trigger mechanism are combined into a new hybrid-trigger mechanism,and the problem of state feedback H_∞control is studied for continuous distributed time-delay systems.In the hybrid-trigger mechanism,the switching of two data transmission modes satisfies the Bernoulli distribution,another Bernoulli variable is used to describe the network attacks phenomenon,and a controller structure containing network spoofing attacks is constructed.Select the appropriate Lyapunov-Krasovskii functional,obtain the sufficient conditions to ensure the mean square asymptotic stability of the closed-loop control system and meet the H_∞performance by random analysis,and obtain the method to obtain the controller parameters by using the holonomic transformation and variable substitution.The effectiveness of the design scheme can be seen through simulation analysis.In addition,the simulation results under different network attacks parameters and different hybrid-trigger parameters are given,and the effectiveness of the design scheme is verified.Finally,the H_∞ control problem of the T-S fuzzy system under a hybrid attack model is studied.It is assumed that there are both deception attacks and replay attacks in the network transmission channel.Using an event-trigger mechanism to save network resources,an event-based hybrid attack model of the networked control system is established.Based on the hybrid attack model,sufficient conditions for the stability of the closed-loop control system are obtained by Lyapunov theory and stochastic theory,and the controller parameters are obtained by LMI technology.The simulation results of two kinds of network attacks with different parameters are compared in the simulation example,and the effectiveness of the design scheme is verified by simulation examples of a spring control system and magnetic suspension control system.