Two‐Step Game‐Based Resilient Model Predictive Control for Cyber‐Physical Systems Against False Data Injection Attacks

ABSTRACT In the paper, a novel resilient model predictive control (MPC) strategy is proposed based on a two‐step game framework for discrete‐time nonlinear cyber‐physical systems (CPSs) under false data injection (FDI) attacks. Firstly, considering the controller's limited capacity to process data at each sampling instant, a two‐step game‐based security framework is designed to adaptively allocate computation and protection resources, ensuring that the transmitted key control input signals can withstand FDI attacks while maximizing the system performance improvement. Secondly, a resilient strategy is proposed to identify the positions of the transmitted key control input signals and further generate a backup control input sequence based on these key control input signals. By combining the game framework with the resilient method, the controller can effectively minimize resource consumption and improve the efficiency of computation resource utilization. Combining these features, a novel resilient MPC algorithm is developed to safeguard the CPS from FDI attacks. Furthermore, the iterative feasibility of the proposed algorithm and the closed‐loop stability of the controlled CPS under FDI attacks are rigorously demonstrated. Finally, the effectiveness of the proposed method is validated through a simulation comparison.