Observer-based Sliding Mode Control For Markov Jump Systems Under Event-triggered Protocol

Environmental mutation and random fault often occur in many practical application fields,such as satellite control,aerospace,power transmission.Markov jump systems is a sort of hybrid stochastic systems containing both modal and state variables,and the existence of Markov chains among subsystems can model the above phenomena well,so it has drawn people's attention extensively in recent years.The phenomenon of network induction follows as the network technology becomes more and more mature.It is well known that the emergence of network induced phenomena can adversely affect the performance of the system,under this case,scholars have proposed several control strategies used to deal with the above phenomenon.Due to the outstanding features such as insensitivity and robustness to dynamic changes,sliding mode control has gradually become one of the most commonly used means.Therefore,the focus of this thesis is that how to decrease the impact of induced phenomenon by sliding mode control method.The main research contents include the following three parts.1.The sliding mode control problem is investigated for uncertain Markov jump systems under event-triggered mechanism.Considering the influence of nonlinearities and delays on system,an appropriate state observer is designed when some states are unknown.An event-triggered scheme is also introduced to reduce unnecessary data transmission.Then the sliding mode dynamics are obtained by the equivalent control method.Subsequently,we provide the sufficient conditions to ensure the fact that the system is stochastic stable by means of Lyapunov technique.Besides,the state trajectory can be driven onto the pre-designed sliding surface in finite time.2.On account of the time-varying delays and nonlinearities,the sliding mode control problem is studied for discrete-time Markov jump systems with packet losses.Moreover,a set of stochastic variables obeying Bernoulli distribution are employed to depict the packet losses case.In addition,to reconstruct the system state,the sliding mode observer has been presented by utilizing the random packet loss probability information.Then,the sufficient conditions are obtained to prove that the system satisfying H? performance is stochastically stable by virtue of the Lyapunov method.Furthermore,the reachability of sliding surface is analyzed by propose a sliding mode control law based on event-triggered protocol.3.Considering the influence of dynamic event-triggered strategy,the sliding mode control problem is studied for discrete-time Markov jump systems under random denial of service attack.In the process of signal transmission,assuming that denial-of-service attacks occur randomly,and the phenomenon of denial-of-service attacks is modeled by Bernoulli random variables.A discrete sliding surface is constructed by using the probability information of random attack and the sliding mode dynamics is obtained by means of the equivalent control method.Then,sufficient conditions are given to prove that the system satisfying H? performance is stochastically stable.In addition,a sliding mode control law is designed based on dynamic event-triggered strategy and the convergence and reachability of sliding surface in finite time are analyzed.At the end of each chapter,the feasibility of the designed control strategy is illustrated through a numerical simulation.