MPC for the Cyber-Physical System with Deception Attacks

In this paper, the model predictive control (MPC) problem is investigated for the cyber-physical system (CPS) with deception attacks. The proposed MPC takes the mixed H
2
/ H
∞
performance index into consideration, by which the attacked system performance can be optimized and the robustness of the CPS can be satisfied. Besides, the cone complementarity linearization (CCL) method is introduced to address the non-convex issue and further the controller gains of the MPC are design. Moreover, the sufficient conditions that guarantee the recursive feasibility of the proposed MPC is demonstrated. Finally, the mass-spring-damping system is used to illustrate the validity of the presented control strategy.