Study On Event-triggered Control And Model-free Control For Switched Linear Systems

The switched system is a typical representative of a hybrid system.The switched system consists of a series of subsystems described by differential and differential equations and a switching signal that determines the switching of these subsystems.Many actual physical systems have this switched characteristic,which gives the switched system a rich application background,and the research of the switched system has also received much attention.With the development of network technology,the scale of control systems is getting larger and larger,and the requirements of control accuracy are becoming higher and higher.In view of the disadvantages of traditional periodic sampling control in terms of information redundancy,a new event-based control method has received widespread attention.As a form of event-triggered control,it has important theoretical value and significance to study the switched system based on dynamic event-triggered control.On the other hand,due to the increasing scale of the system,it is very difficult to obtain accurate system models.The method based on model control is not feasible.It is particularly important and necessary to study control methods that do not rely on system models.In this paper,based on the control theory of the switched system,combined with dynamic event-triggered control and model-free control method,the analysis and design problems of the switched system are studied.The main research contents are as follows:Firstly,the dynamic event-triggered control of continuous-time switched linear systems is studied in two cases,the system state is measurable and unmeasurable.Under the designed dynamic event-triggered conditions,combined with the average dwell time method,a piecewise quadratic Lyapunov function is constructed to obtain sufficient conditions for the stability of the switched system based on the linear matrix inequality,and the subsystem controller design method is also given.The feasibility of this method is verified by numerical examples.Secondly,for the continuous-time switched linear system controlled by the event trigger mechanism,the H_?control problem is studied.Firstly,for the static event triggering mechanism,we obtain the sufficient condition that the switched system is stable and satisfies the H_?performance,and give the design method of subsystem controller based on linear matrix inequality.Then the performance of the switching system under dynamic event-triggered control based on static event-triggered is verified.Finally,simulation examples verify the feasibility of the proposed design method.Finally,assuming that the system model is unknown,the optimal control problem of the switched system under the action of the known switching signal is studied.The data-based Q-learning algorithm is introduced.During the activation period of the switched subsystem,it does not rely on any system model information,only on the collected system state information,and uses reinforcement learning and least squares to solve the subsystem controller online.Finally,a simulation example proves the effectiveness of the proposed algorithm.