Consensus-Based Distributed Robust Filtering for Multisensor Systems With Stochastic Uncertainties

Abstract

This paper proposes a distributed robust Kalman filter for time-varying uncertain linear multisensor systems subjected to stochastic uncertainties. A consensus algorithm is utilized to compromise on a single data among information of all nodes in a multisensor system. Distributed filtering based on consensus remarkably stands out from other algorithms because each node not only estimates its local states correctly, but also reaches an agreement with other nodes over a sensor network as well. The distributed filtering problem is formulated here using consensus on estimation (CE) algorithm. The optimal parameters of the filter are computed by minimizing the covariance of the estimation error. Moreover, the stability of the touched upon algorithm is proved by the Lyapunov stability theorem. Simulation results for a multisensor system with 100 nodes are presented to show the effectiveness and performance of the proposed CE-based distributed robust filtering approach.