Road-Map Aided VSIMM-GMPHD Filter for Ground Moving Target Tracking
Jihong Zheng, Meiguo Gao, Haojie Yu · 2018
Tracking multiple ground moving targets with uncertain target dynamics proves to be a complicated task, especially when the number of targets varies. This paper presents a road-map aided variable structure interacting multiple model Gaussian mixture probability hypothesis density (RA-VSIMM-GMPHD) filter for ground moving targets tracking using moving target detector (MTD) measurements obtained from a monostatic radar. The probability hypothesis density (PHD) filter, which is a recursive Bayesian algorithm for jointly estimating the time-varying number of targets and their states in clutter, has been shown to be a computationally efficient solution to multi-target tracking problems with a varying number of targets. Specifically, the Gaussian mixture PHD (GMPHD) is a closed-form solution to the PHD recursion under linear, Gaussian assumptions on the target dynamics. The target dynamics uncertainty can be resolved by road-map aided variable structure interacting multiple model (VSIMM) estimator, in which the mode sets are modified adaptively depending on the road-map. The RA-VSIMM-GMPHD filter, which incorporates road-map aided VSIMM estimator into a framework of GMPHD filter, is particularly suited to tracking ground moving targets with uncertain dynamics and number varies. Simulation results are presented to show the effectiveness of the proposed filter over single-model GMPHD filter and VSIMM estimator.