PHD filter with amplitude information in weibull clutter

Abstract

Tracking performance can be considerably improved if amplitude information (AI) is used, since AI helps to obtain more accurate target and clutter likelihoods. The Probability Hypothesis Density (PHD) filter using AI has been studied in Raleigh distributed clutter. However, Raleigh distribution is not suitable when describing the heavy-tailed clutters observed by high resolution radar or when grazing angle is small. In this paper, we consider the multi-target tracking in Weibull clutter in the framework of PHD filter. The closed form solution of PHD filter is derived based on uniform target birth model for situations where the signal-to-clutter ratio (SCR) is both known and unknown. Simulation results for Gaussian mixture implementation of PHD filter show that significant performance gain can be obtained when tracking targets in Weibull clutter.