Optimal Security Control for Cyber-Physical Systems Under Cyber-Attacks

In this paper, the research focuses on the security control issues of cyber-physical systems (CPSs) subjected to stochastic denial of service (DoS) attacks. Consideration is given to the occurrence of stochastic DoS attacks within the communication network from the controller to the actuator, which can lead to network congestion, impeding signal transmission to the actuator, thus affecting the secure operation of CPSs. Bellman dynamic programming techniques are employed in this paper to derive optimal control gains along with a corresponding algebraic Riccati equation. Subsequently, a Q-learning algorithm is utilized to achieve data-driven controller design objectives. Finally, simulation cases are conducted to validate the effectiveness of the security control approach proposed in this paper.