Fault detection for descriptor systems with Markov jump parameters and time-varying delays

This paper investigates the observer-based fault detection filtering problem for continuous-time descriptor Markov jump linear systems (DMJLSs). Both parameter uncertainties and state delays are assumed to be time-varying but bounded. A fault detection filter (FDF) is designed such that for all admissible uncertainties, the resultant filtering dynamics is robust stochastically stable, and the residual signal is robust to the unknown input and the known control input but sensitive to the fault. Moreover, an iterative linear matrix inequality (LMI) approach for filter parameter computation as well as the residual evaluation and threshold calculation are also given. The validity of the proposed method is illustrated by a numerical example.