Distributed Filter Based on SICI Data Compression over Sensor Networks

The distributed filtering problem based on data compression is studied for linear discrete time-varying systems over sensor networks. At each sensor node, sequential inverse covariance intersection (SICI) fusion algorithm is used to compress the estimates received from neighbor nodes. A distributed recursive filter is presented based on compressed data. The filtering gain (FG) and consensus gain (CG) are solved by locally minimizing an upper bound of the filtering error covariance matrix (FECM). It avoids the calculation of cross-covariance matrices (CCMs) and reduces the computational burden. Simulation results verify the effectiveness of the algorithm.