Improved Probabilistic Multi-hypothesis Tracker for Multiple Targets Tracking with Discrete Feature

In tracking scenarios of high resolution radar, it is possible to obtain more information about targets. This paper demonstrates how the Probabilistic Multi-Hypothesis Tracker (PMHT) can be extended to include discrete feature information when both the uncertainty of feature models and the instability of feature observing process should be taken into consideration. A framework for multiple targets tracking with discrete feature measurements is presented based on a probabilistic integration of discrete feature state estimation and tracking process. A Monte Carlo simulation study has been employed to identify the target tracking where performance improvement is obtained over the standard PMHT and the PMHT-C.