Data association based on target topology

Data association is a key problem in multiple-sensor multiple-target tracking. Normally, there is bias of location while the platform of sensor is moving. In that case, the measurement for the same target from multiple sensors will have a large bias through coordinate transforming. It is a boring problem for utilizing position and velocity of target to associate. In order to solve this problem, topological information of targets was abstracted. And based on them, through matching and searching, an optimal topological pair was obtained. Then this topological pair was provided to prior information for data association and marching. The simulations show that this method has better accuracy and robusticity.