Skip to Main Content
Based upon the graph searching method in ordered state space of combinatorial problem, we define the joint event as the problem node and transform a problem of data association for multitarget tracking to one in state space. In order to reduce the number of extended nodes during searching, we define the function of the joint probability data association as the valuation function of node and present a new simple approach to determining whether a node is an old one produced previously. After all we put forward a new algorithm which can easily search out N events with the largest joint probabilities and calculate their joint probabilities quickly. The new algorithm has the advantages of both Bayesian and non-Bayesian methods.