Several methods of radar data fusion

In the processing of radar data, target estimation states fusion is an important problem. Especially under the decentralized model, various estimations of target state which come from local filters with Kalman filter may be fused in order to obtain a more accurate target estimation state. There are two problems to be solved. The first is the fusion criterion, different fusion criteria bring out different results. Three fusion criteria are discussed; one is based on the covariance matrix of estimation error, the second is dependent on the infinite normal number, and the last is related to the element of the state vector. The second problem is the tracking algorithm; it affects the accuracy of the state fusion result. The three criteria and their corresponding algorithms are presented. The algorithms estimating target state are set up with a strong tracking filter other than Kalman or extended Kalman filter; they have better tracking performance so that they can overcome the deficiencies related to Kalman or extended Kalman filters. The reason for this is explained. Comparison results are given.