For the problem of tracking multiple targets, the joint probabilistic data association (JPDA) approach has shown to be very effective in handling clutter and missed detections. The JPDA, however, tends to coalesce neighboring tracks. In this paper the authors develop probabilistic filters that avoid the JPDA’s sensitivity to track coalescence and preserve the resistance to clutter and missed detections.
At the beginning, a short introduction to multi-target tracking problems is given and the differences of various known methods to approach such problems are discussed. Then the authors embed the multi-target tracking problem into one of filtering given measurements from a linear descriptor system with stochastic coefficients and develop various filter algorithms. The results are illustrated by Monte Carlo simulations.