Efficient multi-target tracking with sub-event IMM-JPDA and one-point prime initialization

This paper addresses an IMM (interacting multimodel)-JPDA (joint probabilistic data association) tracker with sub-event decomposition and one-point prime initialization. In the original JPDA, the number of joint feasible association event is about the factorial size of the target numbers, which may cause the huge computational burden. The proposed sub-event JPDA can significantly reduce the number of hypotheses without reducing tracking performance. One-point prime initialization method estimates initial velocities using initial range rates and azimuths. A scenario of 10 multiple targets is tested and the performance is evaluated in terms of root mean-squared errors of position and velocity and the averaged mode probabilities. The computational time of the sub-event IMMJPDA tracker is compared with that of the IMM-JPDA showing the significantly improved time efficiency of the proposed method.