Abstract:

A generalized distributed track-to-track  association method is proposed for multiple targets tracking in collaborative sensor network. Firstly, by analyzing the traditional sequence  track-to-track  association rules, all pairwise likelihood function and generalized  track-to-track  association statistic of the distributed tracking are constructed. On the basis of that, the mathematical model of the generalized track-to-track  association is established. Then the distributed  track-to-track  association question is translated into the multiple dimension assignment. Secondly, the optimal association relationship among tracks is searched by using the improved  immunity algorithm. The whole  track-to-track  association course includes two stages. At first, the unrelated track combination is eliminated by using the chisquare distribution hypothesis test, and then the multiple sensors multiple tracks association relationship is confirmed according to the immunity antibody fitness. Finally, the proposed method is evaluated by utilizing some track-to-track  association performance parameters. Simulation results show that the association stability and the association effect of the new method are improved obviously compared with the traditional sequence pairwise association method.