Asynchronous fault detection filtering for piecewise homogenous Markov jump linear systems via a dual hidden Markov model.

• Design an asynchronous fault detection filter for non-homogeneous MJSs. • The system modes can be traced by an asynchronous fault detection filter. • Sufficient conditions are devised to ensure the designed system performance. • The AETB is used to validate the devised asynchronous fault detection filter. This work investigates the asynchronous fault detection filtering issue for discrete-time piecewise homogeneous Markov jump systems. By resorting to a dual hidden Markov model, an asynchronous fault detection filter is presented which can follow up the system modes. Utilizing the stochastic Lyapunov function approach, sufficient conditions are devised to make the resultant Markov jump systems be stochastically stable and hold a specified H ∞ performance level. Finally, a real-time experiment based on the automotive electronic throttle body is employed to validate the devised asynchronous fault detection filter which can find the faults after the occurrences in the case of no false alarm. [ABSTRACT FROM AUTHOR]