Dynamic model predictive control for constrained cyber-physical systems subject to actuator attacks

Modern networked control systems are cyber-physical systems (CPSs), in which the cyber space is tightly integrated with the physical world and human intervention. Forming the basis of future smart services, CPSs are expected to significantly promote our life. However, CPSs also suffer from cyber-attacks due to their increasing connections to the Internet. This paper investigates secure model predictive control (MPC) for constrained CPSs subject to actuator attacks, which intentionally manipulate control commands from a controller to actuators. In our study, the theory of invariant set is used to construct a Luenberger observer for error-bounded state estimation in the presence of actuator attacks. Then, a robust output feedback MPC controller is designed for constrained and attacked systems based on a set-membership state estimator. After that, stability conditions are theoretically established for the proposed MPC controller. A numerical example is also given to demonstrate the effectiveness of the proposed secure MPC approach.