Delayed Merging Gaussian Mixture PHD Tracker with Embedded MHT for Close Target Tracking

摘要

For multi-target tracking it is difficult to obtain target trajectories from measurements of close targets and clutter. The Gaussian Mixture Probability Hypothesis Density (GMPHD) as a closed form solution for the Probability Hypothesis Density (PHD) filter can easily provide track labels of targets in clutter. But when targets are too close to each other, such as crossing and occluded conditions, the GMPHD tracker can?t resolve identities of the targets which affects and even interferes with the decision of the commander. Based on the separation distance we proposed, delayed merging GMPHD tracker is proposed to correctly track close targets in clutter. Simulation results show that our proposed approach significantly improves the tracking performance of the GMPHD filter for correctly identifying targets in close proximity.