A greedy assignment algorithm and its performance evaluation

Data association is a critical problem in multitarget tracking. In fact, it is the bottleneck of most of the multitarget tracking algorithms such as multiple hypothesis tracker (MHT). In this paper, we propose a heuristic assignment algorithm which determine approximately the N best data association hypotheses in an efficient manner. This "greedy" algorithm is based on the "branch and bound" concept and has the flexibility of adapting to various scenarios. Performance evaluation of the algorithm based on simulation is also presented.