Event-triggered stochastic model predictive control based on distributionally robust optimization approach for network control systems under DoS attacks

Abstract

In this paper, we propose a novel event-triggered stochastic model predictive control (SMPC) algorithm based on distributionally robust optimization (DRO) for network control systems (NCSs) that are affected by additive disturbances and Denial-of-Service (DoS) attacks. The event-triggered mechanism is designed to reduce sampling times, communication costs, and computational demands while maintaining performance. By converting chance constraints into second-order cone (SOC) constraints, the algorithm effectively addresses disturbances with limited prior knowledge, utilizing accessible first and second moments. The proposed event-triggered DR-SMPC (ETM-DR-SMPC) has longer sampling intervals than periodic sampling DR-SMPC (PDR-SMPC). Compared to self-triggered SMPC (SSMPC), ETM-DR-SMPC has longer sampling intervals, less computation, better performance and fewer constraint violations. It also has better performance and longer sampling intervals than event-triggered robust model predictive control (ETM-RMPC). Furthermore, we demonstrate the exponential mean square stability of ETM-DR-SMPC for the first time and the recursive feasibility of ETM-DR-SMPC is guaranteed. Numerical simulations validate the effectiveness of the proposed algorithm.