Stackelberg Game-Based Resilient MPC for Nonlinear CPSs under FDI Attacks

This article proposes a Stackelberg game-based resilient model predictive control (MPC) for a constrained discrete-time cyber-physical system (CPS) against false data injection (FDI) attacks. First, a Stackelberg game-based resilient method is designed to determine and protect the important control samples within MPC optimal control data package, based on which an alternative feasible control sequence could be generated even if the original one is tampered by the FDI attacks. Then, a novel resilient MPC algorithm is proposed to ensure the safety and smooth operation of the CPS under FDI attacks and further reduces the consumption of the system resources. Moreover, the feasibility of the proposed MPC algorithm and the closed-loop stability of the CPS are rigorously demonstrated via theoretical analysis. Finally, it is verified through robot experimentation that the designed algorithm can effectively alleviate the malicious influence of FDI attacks while reduce the controller's resource consumption.