Robust Control And Synthesis Of Event-driven Discrete-time Network Systems

We first focus on the robust and resilient finite-time guaranteed cost control problem of discrete network systems with event-triggered mechanisms;subsequently,the design of event-triggered and quantized finite-time synchronization of discrete neural networks with Markov jumps is studied;finally,event-triggered finite-time quantized control is investigated for discrete-time singular Markov jump systems.The main work of the paper is as follows:1)The problem of event-triggered robust finite-time guaranteed cost control is studied for discrete uncertain network systems.By introducing a novel event-generator in the control loop systems,an error system model is proposed,and the definitions of finite-time boundedness and robust finite-time guaranteed cost control are presented.Next,with the help of Lyapunov functional and free-weighting matrix techniques,criteria on finite-time boundedness and finite-time guaranteed cost boundedness are derived for the discrete network model.A resilient robust finite time controller and event-triggered parameter are co-designed to ensure the finite time boundedness of the class of network model.Finally,the effectiveness of the proposed method is illustrated by a numerical example.2)Finite-time synchronization problem is investigated for event-driven and quantized discrete Markov neural networks.Firstly,a unified framework based on event-triggering schemes and quantization,a synchronization error model is constructed for discrete Markovian neural networks.Secondly,with the aid of Lyapunov functional,sufficient conditions for the finite-time boundedness are deduced which can ensure synchronization of the system,and the quantized finite-time controller and event-triggered parameter are co-designed to guarantee the finite-time boundedness of the Markovian neural networks by variable decoupled techniques.3)This project concerns the singular finite-time bounded H_∞admissible problem of discrete-time singular Markov jump network systems with event-triggered mechanism and quantizations.According to event-triggering communication strategy and quantized control mechanism,a singular Markov jump network systems model is initially derived.Then,sufficient criteria are established to ensure that the singular Markov jump network systems is singular finite-time bounded H_∞admissible.Furthermore,event-triggering finite time quantized controllers and event-triggering matrix gains are co-designed for the singular Markov jump network systems model based on variable decoupling method,and the feasibility of proposed approach can be converted into convex optimization problem.