Labeled particle unresolved target PHD filter for multiple group target tracking

When the sensor is incapable of continuously achieving measurements for each target for close target distance or low sensor resolution, such targets is usually treated as a group target, and multiple group target tracking is a special case of multiple target tracking. On the basis of unresolved target probability hypothesis density (UT-PHD) filter, Sequence Monte Carlo (SMC) method and labeled track extraction method, labeled particle UT-PHD (L-PUT-PHD) filter is proposed in this paper for estimation of group target number as well as group center state and track extraction, realizing detection and tracking for multiple group targets. Simulation experiments show that L-PUT-PHD filter performs well on detecting appearance and disappearance group target, estimating location of the group center as well as extracting track of group target.