Distributed stubborn-set-membership filtering with a dynamic event-based scheme: The Takagi-Sugeno fuzzy framework

The article develops a novel stubborn set-membership filtering with the dynamic event-triggered scheme for discrete-time nonlinear systems with the state-delays, unknown-but-bounded noises as well as abnormal measurements. First, in comparison with traditional event schemes, a dynamic event-triggered scheme with a time-varying auxiliary offset variable in the threshold is employed to schedule the access token with the purpose of mitigating the communication burden. A novel fuzzy stubborn filter in a distributed way is then constructed to obtain the ellipsoid set including the real state while constraining the direct influence of abnormal measurements, which could come from outliers or a malicious modification by attackers. An algorithm with the form of recursive linear matrix inequalities is presented to determine the desired ellipsoids as well as the distributed filter gains, where the operation of the algorithm do not depend on the global information of network topology. Finally, the validity of the proposed scheme is illustrated via an inverted pendulum system.

Automated summary

The article presents a novel stubborn set-membership filtering approach for discrete-time nonlinear systems with state-delays, unknown-but-bounded noises, and abnormal measurements. By using a dynamic event-triggered scheme and distributed fuzzy filter, the real state is obtained while mitigating the impact of abnormal measurements. The proposed algorithm for determining the ellipsoids and filter gains operates independently of global network topology information.