Event-Triggered Predictive Control for Semi-Markov Jumping Networked Control Systems With Random DoS Attack

Networked control systems (NCSs) are an important component of industrial cyber-physical systems (ICPSs). This paper proposes an event-triggered discrete-time stochastic semi-Markov hopping networked control systems robust predictive control strategy for ICPSs to address common parameter uncertainties and external disturbances. The control scheme aims to tackle the issues of randomness and time-varying characteristics inherent in ICPS, such as random time delays and denial-of-services (DoS) attacks in the channels of sensor-to-controller (S-C) and controller-to-actuator (C-A). To overcome the inherent bandwidth limitations, a dynamic event-triggered scheme is suggested to alleviate the network strain. The stability analysis gives a sufficient criterion for the $\sigma$-error mean-square stability ($\sigma$-MSS) of a semi-Markov jump linear system (s-MJLS) using Lyapunov function. Based on this analysis, a network predictive controller is designed to guarantee the robustness of the system. A linear switched reluctance motor (LSRM) system, as a typical application of ICPSs, is used for simulation and experimentation, thereby validating the practical efficacy and suitability of this approach.