Distributed Kalman Predictor with Different Consensus Gains over Sensor Networks

This paper presents a distributed Kalman predictor (DKP) with different consensus gains for sensor networks, i.e., a prior filter, only considering a locally minimum upper bound of the prediction error covariance matrix, where each sensor exchanges local estimates with its neighbor nodes. Since calculations of prediction error cross-covariance matrices between sensor nodes are avoided, the proposed DKP can effectively reduce computational cost and be more suitable for distributed way. To obtain the DKP, an optimal Kalman predictor gain for each sensor node and different optimal consensus gains for state estimates of its neighbor nodes are designed to minimize locally an upper bound of the prediction error covariance matrix in the linear unbiased minimum variance (LUMV) sense. Stability and steady-state property of the DKP are analyzed. An example for vehicle tracking in sensor networks demonstrates effectiveness of the proposed algorithm.