A distributed fault estimation method for a class of nonlinear Markov jump systems

Abstract

This paper investigates the distributed fault estimation problem of a class of nonlinear Markov jump systems. The main goal is to design a distributed observer network to estimate system states and actuator fault simultaneously. For the two cases of completely known and general uncertain transition rates, we first use a virtual intermediate observer technique to remove the observer matching condition, which upgrades the flexibility in the process of design. Next, by using the information of some special neighboring nodes and adopting observability decomposition approach, a novel distributed observer network allowed to be with unobservable nodes is constructed. Then, the mentioned goal is achieved based on the observer network. The proposed method can compensate missing information of unobservable nodes, which solves the estimation problem in the case of partial information loss. Finally, an example is provided to demonstrate the effectiveness of the proposed method.