Event-based Distributed Filtering Approach to Nonlinear Stochastic Systems over Sensor Networks

Abstract

In this paper, an event-triggered communication strategy and a distributed filtering scheme are designed for discrete-time nonlinear stochastic systems over wireless sensor networks (WSNs). The underlying system is represented by the Takagi-Sugeno (T-S) fuzzy model, and in addition by the description of the WSN under consideration. The structure of the WSN is established on a deterministic one. Based on an event-triggering condition tailored for each sensor, distributed fuzzy filters are established using the triggered measurements of the smart sensors. As a result, an augmented stochastic system is presented for the distributed filtering design. A robust mean-square asymptotic stability criterion is explored using the Lyapunov stability theory and the Disk stability constraint is applied to improve the performance of the distributed filters. An optimization solution to obtaining the parameters of the distributed filters is developed. Subsequently, a computer-simulated example helps to illustrate the validity of the proposed new filtering design techniques.