Distributed estimation in general directed sensor networks based on batch covariance intersection
Tao Sun, Ming Xin, Bin Jia · 2016
The paper presents a distributed estimation scheme based on a new batch covariance intersection (BCI) strategy and an average consensus algorithm to address the problem of data fusion in sensor networks. Due to sharing common prior knowledge, process noise and/or existence of correlated measurement noise, the error of the local estimates from each sensor node in a sensor network is correlated with each other to some extent with unknown cross-correlation. The BCI scheme can handle the correlation in the data fusion in a distributed way by means of an average consensus algorithm so that no fusion center is needed. Moreover, the proposed average consensus algorithm can be applied in a general digraph including the non-balanced topology. A cooperative target tracking problem using multiple UAVs as the mobile sensor network is used to demonstrate the performance of this new distributed estimation algorithm.