Track Initiation Technique and Its Application in PHD filter

Abstract

The probability hypothesis density (PHD) filter has been proved to be an effective method for multi-target tracking in the presence of false alarms, missed detections and an unknown numbers. However, the PHD filter fails to obtain the state estimation of a target accurately during its initial time. This will degrade the performance of the PHD filter on the detection of new targets. In order to resolve this problem, we apply the rule-based track initiation technique to the Gaussian mixture PHD filter, and propose the Gaussian mixture PHD filter with track initiation. The results of the simulation experiment demonstrate that the proposed filter can achieve better performance on tracking multi-target than the Gaussian mixture PHD filter.