Fault Detection Filtering for a Class of Nonhomogeneous Markov Jump Systems with Random Sensor Saturations

Abstract

This paper is concerned with the problem of fault detection filtering for a class of nonhomogeneous Markov jump systems with randomly occurring sensor saturations, where it is assumed that sensor saturation is randomly occur satisfying the Bernoulli distribution, meanwhile transition probabilities in nonhomogeneous Markov chain are time-varying. By employing the suitable Lyapunov technique in combination with Linear Matrix Inequalities (LMIs), some sufficient criteria are derived to guarantee that the augmented system is mean square stochastically stable with an H∞ performance level. Finally, a representative example of industrial stirred tank reactor systems is presented to illustrate the theoretical results, and it then turns out that the feasible filtering parameters can be exactly fixed by Matlab Tool Box, and the H∞ performance of the system can be fully guaranteed by the proposed control strategy.