Fuzzy-Model-Based Asynchronous Fault Detection for Markov Jump Systems With Partially Unknown Transition Probabilities: An Adaptive Event-Triggered Approach

Abstract

This article addresses the event-triggered asynchronous fault detection (FD) problem of fuzzy-model-based nonlinear Markov jump systems (MJSs) with partially unknown transition probabilities. For this objective, the nonlinear plant is modeled as an interval type-2 (IT2) fuzzy MJS with the aid of the IT2 fuzzy sets capturing the uncertainties of the membership functions. An adaptive event-triggered scheme is introduced to bring down the costs of the communication network from the system to the fuzzy fault detection filter (FDF), in which the triggering parameter can be adaptively tuned with the system dynamics. A hidden Markov model (HMM) is employed to characterize the asynchronous phenomenon between the system and the FDF. Unlike the existing results, the transition probabilities of the plant and the FDF are allowed to be partially known. By using the Lyapunov and the membership-function-dependent methods, the existence conditions of the FDF are derived. Finally, the proposed FD methods are verified by a numerical simulation.