Abstract:

In multitarget tracking in clutter, often there are more than concerned measurement available for updating the state of a signal target. To solve the problem of data association between targets and measurements, a fast algorithm is proposed for data association in dense clutter based on the principle of joint probabilistic data association(JPDA). This algorithm is more efficient to directly compute a posterior probabilities without generating the data association hypotheses. The simulation results show that the computational cost for data association of the fast algorithm is significantly reduced than that of JPDA, and the tracking performance is very good in the case of more than 100 targets in dense clutter.

Key words: multiple target tr acking, dat a association, r adar tr acking