Neural network-based dynamic output feedback control for nonhomogeneous Markov switching systems under deception attacks

This paper presents a neural network-based method to address the challenge of designing dynamic output feedback controllers for nonhomogeneous Markov switching systems (NMSSs) under deception attacks. The model enhances realism by incorporating a nonhomogeneous Markov process to depict the system's stochastic switching behavior. To alleviate communication load and prevent frequent data collisions, a round-robin protocol is implemented for transmitting measurement outputs. Unlike conventional approaches that assume deception attacks are known and bounded, this work considers more general unbounded deception attacks and employs neural networks to approximate and mitigate their impact on the system. Utilizing Lyapunov stability theory, sufficient conditions are derived to ensure the stochastic stability of the closed-loop system. Finally, the effectiveness of the proposed approach and the theoretical results are demonstrated through a simulation example.