Asynchronous fault detection for hidden Markov jump systems with dynamic event-triggered mechanism

In the present paper, the problem of asynchronous fault detection (FD) for Markov jump system based on a dynamic event-triggered transmission mechanism has been discussed. Noting that the system modes cannot be accurately obtained due to the common network induced factors in practice, a hidden Markov model (HMM) is introduced to describe the asynchronous behavior between the system mode and the one of filter. To further save the limited network bandwidth as well as the computational resources, a novel dynamic event-triggered transmission mechanism (DETM) is presented via constructing a special matrix threshold, in which the entries can be flexibly adjusted to meet different performance requirements of each subsystems. Then, a mode-dependent Lyapunov function is constructed to acquire a sufficient condition, which enables the system stochastically stable with H_∞ performance. And some criteria with less conservatism are also developed by means of the Finsler’s lemma, such that the filtering performance can be further improved. Lastly, the feasibility and practicality of the event-based FD filter are verified by an example simulation.