Distributed Kalman Filtering based on Optimal Weighting Matrix

In this paper, we consider the fusion problem of distributed Kalman filtering over a multi-sensor network, where each sensor can exchange information only with other sensors within its transmission/reception range through single-hop communications, and each sensor makes an estimation for the current time state of linear discrete-time dynamic system. We propose a distributed algorithmic framework for finding the optimal weighting matrices analytically, which leads to the least MSE at each sensor. Finally, two numerical simulations are used to illustrate the fact that the proposed distributed filter design scheme has a better performance than the existing distributed method. Consequently, it is more close to the centralized Kalman filtering.