Distributed filtering for a class of discrete-time systems over wireless sensor networks

This paper addresses the distributed filter design problem for a class of dynamic systems over wireless sensor networks. The missing measurements and the correlation among state noises and measurement noises are considered, where a set of mutually uncorrelated random variables is employed to describe the missing phenomena. Firstly, the construction of the designed filter is proposed and the prediction of the state at each node is given. Then, the filtering error covariance is presented and the filter parameters are determined to minimize the trace of such a covariance, where the network topology data are used to simplified the singular matrix. Subsequently, the relationship between the filter performance and missing probability of the measurement is discussed. Finally, a numerical simulation is presented to illustrate the effectiveness and capability of the proposed distributed filters.