Distributed Filtering with Communication Constraints, Randomly Occurring Nonlinearities and Correlated Noises Over Sensor Networks

In this paper, the distributed recursive filtering problem is addressed for discrete time-varying systems with correlated noises, randomly occurring nonlinearities and communication constraints. The communication constraints refer to the maximum-error-first (MEF) protocol, which are utilized to save the communication resources and reduce the communication burden. The main purpose of this paper is to design the optimal distributed filter under minimum variance sense for the addressed discrete time-varying systems, where a recursive distributed filtering algorithm is proposed. In particular, an upper bound of the filtering error covariance is given by solving the Riccati-like difference equation, and the filter gain is designed to minimize the upper bound of filtering error covariance. In addition, a new matrix simplification technique is introduced to deal with the sparsity of the sensor networks. Finally, the applicability of the recursive distributed filtering algorithm is illustrated by simulation.