Multitarget tracking using dominant probability data association

A new suboptimal approach to the probability data association of multitarget tracking, the Dominant Probability Data Association (DPDA), is presented in this paper. In view of the fact that the case where many targets cross together or move in a very 'small' neighborhood, rarely occurs for most practical multitarget tracking environments we may define a dominant joint event and corresponding dominant joint probability. Using Bayesian rule, we can deduce a formula of the dominant joint probabilities without calculating the all joint probabilities of all joint events such as in joint probability data association (JPDA). So, the DPDA can avoid the problem of combinatorial 'explosion' in JPDA. In addition, we prove that the top limit of performance of DPDA is equal to that of JPDA and the low limit is not lower than that of probability data association (PDA) and that the event with low limit is the one with very small probability. Monte Carlo simulation results give out inspiring performance.