Protocol-Based Control for Hidden Markov Jump Systems With Incomplete Transition Descriptions Against Injection Attacks
Yuan Wang, Huaicheng Yan, Ju H. Park, Hao Zhang, Hao Yang Shen · IEEE Transactions on Control of Network Systems · 2024
This article aims to design the dynamic event-triggered protocol for hidden Markov jump systems subject to injection attacks. We initially introduce the hidden Markov model (HMM), which accommodates incomplete transition descriptions found in both the transition probability matrix and the observation probability matrix of the observation signal. Importantly, the underlying HMM demonstrates applicability to a broader range of scenarios than certain existing works. Furthermore, data are vulnerable to cyberattacks before being transmitted to the controller due to the open communication architecture, in which adversaries are assumed to inject false data that are capable of affecting data security. To effectively mitigate unnecessary utilization, we develop a novel dynamic event-based protocol involving dynamic threshold parameter and system modes in the feedback channel. Our methodology tackles the vulnerability of data to cyberattacks in open communication architectures, and the proposed protocol provides a feasible security solution for the hidden Markov jump systems. Eventually, the theoretical findings are demonstrated by two provided simulation examples.