An efficient algorithm for data association in multitarget tracking

An efficient algorithm is developed to compute the a posteriori probabilities of the origins of measurements in the joint probabilistic data association filter (JPDAF). The inherited parallelism of this algorithm enables it to be suitable for implementation in a multiprocessor system. In this algorithm the a posteriori probability of the origin of each measurement in the JPDAF is decomposed into two parts. The computation of one part becomes trivial and the algorithm developed here is implemented on the other part, which is shown to be related to permanents. The computational complexity of this algorithm is analyzed in the worst case as well as in the average case. An analysis of this algorithm enables us to conclude that this algorithm is more efficient than other existing ones in the average case.<>