A new decentralized data fusion algorithm with feedback framework based on the covariance intersection method

The main objective of this paper is to analyze models, techniques, algorithms and infrastructures needed to complete decentralized data fusion. It is often to use the Simple Tracking Fusion Algorithm in decentralized data fusion for Multi-sensor. But that algorithm is on the hypothesis that the output of each local filter is uncorrelated. If the fusion result is given back to each local filter, the output of the local filter will has correlation with each other. For that case, if still using the Simple Tracking Fusion Algorithm, the fusion data will lose consistency. But in the Covariance Intersection Algorithm, it is unnecessary to consider the correlation of each local filter. In this paper, a new decentralized data fusion algorithm with feedback framework based on the Covariance Intersection algorithm is proposed for Multi-sensor. The simulation results show the effectiveness and robustness of the proposed algorithm.