RFS-Based Multiple Extended Target Tracking With Resolved Multipath Detections in Clutter

In the literature, the problem of point target tracking with multipath detections has been studied. However, the case of extended target tracking in a multipath environment (e.g., tracking a submarine using a high resolution sonar, tracking a vehicle in an urban environment using an imaging radar) has not been adequately addressed. If the multipath detections from a single target can be modeled and used properly, better tracking performance can be obtained in terms of accuracy, false tracks and computing time. By integrating the Random Matrix (RM) theory and the random finite set (RFS) theory, an extension of the Probability Hypothesis Density (PHD) filter, called MP-ET-PHD, is proposed in this paper to address the multitarget tracking problem with an unknown number of targets in an uncertain multipath environment with clutter. In the proposed framework, a novel multipath measurement update equation is formulated and derived. Also, a Gaussian Mixture (GM) implementation of the proposed MP-ET-PHD is presented for practical applications. Simulation results show that the proposed MP-ET-PHD can effectively handle multipath detections and yield improved tracking performance over the traditional single-path extended target trackers.