State estimation and data fusion in a distributed monitoring system network

The aim of this paper is to propose a method to estimate states of a target considering observations of multi monitoring systems with different probability of detection. The proposed scenario includes a collection of processing nodes or sensor nodes in a distributed network. Joint target detection and tracking are done in the processing nodes by the particle filtering conditioned on the observation of its own sensors. Then the estimated states are fused by received estimation of other nodes based on covariance intersection. A modified covariance intersection is proposed in this paper which reduces the mean square error and covariance of estimation. The simulation results illustrates that mean square error for the fused estimation are less than the state estimations in each processing nodes. The results is also compared with Kalman filter and particle filter.

Keywords:

state estimation , target detection and tracking , particle filtering , covariance intersection