Track correlation for multisensor system based on NN and general classical assignment algorithm

Track correlation is the key problem in the information fusion of the distributed multisensor system. The purpose is to judge whether or not the tracks In a distributed multisensor system, coming from different local nodes are maed to a same target. The flow of the track corelation of multisensor system is proposed based on the analysis about the problem. At first, space and time registration for the datum detected by different kinds of sensor from different place were taken with NN technology. Then the problem of track correlation were transformed to the generalized classical assignment problem and NN were used to deal with this problem. The simulation experimental results illustrate that with above method, the system has higher correct association rate.