Robust Reconstruction-Based Resilient MPC for Industrial Cyber-Physical Systems Under Deception Attacks

Abstract

In the article, we propose a novel robust reconstruction-based resilient model predictive control (RR-MPC) strategy to protect industrial cyber-physical systems (ICPSs) against deception attacks. To reach this goal, a robust reconstruction strategy is first designed, which could not only predict the additional disturbance of the ICPSs to achieve a less conservative state error but also determine the specific key protected input samples to calculate the feasible ones if deception attacks tamper with the optimal one. Based on the robust reconstruction strategy, a novel resilient MPC algorithm is designed to maintain the system operation stability under deception attacks while also reducing the controller’s computing and security resources consumption. Moreover, the recursive feasibility of the modified MPC algorithm and the closed-loop stability of the ICPSs driven by the considered algorithm are all demonstrated. Finally, the effectiveness of the proposed resilient MPC algorithm is verified via a simulation example and a robot-based experimental verification.