A Multiple Model GM-PHD Filter for Maneuvering Target Tracking.docVIP

A Multiple Model GM-PHD Filter for Maneuvering Target Tracking Xiao Wang #1, Chongzhao Han #2 # School of Electronic and Information Engineering Xi’an Jiaotong University Xi’an, Shaanxi, China 1 wangxiaox@ 2 czhan@ Abstract—PHD is a novel approach to multi-target, multi-sensor tracking based on finite random set theory. The advantage of PHD is that it can deal with targets with unknown number. But for maneuvering target it does not have special advantage. In this paper, a implementation of multiple model Gaussian Mixture PHD filter has been proposed. It has the advantage of both PHD filter and multiple model method and can deal with multi-targets’ maneuvering. Theoretical analysis and simulation in different scenarios have showed the algorithm’s rationality and effectiveness. Keywords: Maneuvering target racking, probability hypothesis density, multiple model, estimation. Introduction The probability hypothesis density (PHD), popularized by Ronald Mahler, is a novel approach to multi-target, multi-sensor tracking. Based on finite random set theory, the PHD is the first moment of a point process of a random track set, and it can be propagated by Bayesian prediction and observation equations to form a multi-target, multi-sensor tracking filter. The PHD is the factorial moment density found in point process theory, and it provides a straightforward method of estimating the number of targets in the region under observation. The PHD filter propagates the PHD or the first moment of the multi-target posterior density where the integral of the PHD over a region in the state space is the expected number of targets within this region, and the peaks in the PHD can be regarded as the estimated locations of the targets at a given time step. Sequential Monte Carlo PHD is an implementation of PHD filter [1]. The main drawbacks of the approach are the large number of particles and the unreliability of clustering techniques for extracting state estimates. Vo proposed a closed-fo