Learning Optimal DoS Attack Scheduling for Remote State Estimation Under Uncertain Channel Conditions

Recently, the security of cyber-physical systems is paid more attention gradually. In this paper, we consider the optimal denial-of-service attack scheduling problems under uncertain channel conditions and the security issues of cyber-physical systems are analyzed from the perspective of attackers. The goal of attackers is to design an attack scheduling to maximize the linear cost function while maintaining the stability of systems. To solve this scheduling problem, the Markov decision process is formulated. Since the channel parameters are unknown, the Q-learning algorithm is proposed to solve the associated optimality Bellman equations. Some simulation results are presented to show the effectiveness of the obtained results.