Feature Aided Extended Target Tracking For High Resolution Radar
Ming Liu, Kai Zhao, Ya Zhang, Yiyue Gao, Tao Zhang · 2021
In high-resolution radar applications, the scatter distribution of target usually evolves non-linearly with the target movement, and thus it is difficult for a Bayesian filter to achieve an accurate estimation of the target extension state on its own. To solve this problem, we propose to use measurement features to help to propagate the posterior of the target extension in this paper, resulting in a feature-aided extended target probability hypothesis density (FA-ET-PHD) filter, where the features are applied to calculate the partition weights. Since the feature-based weights do not abide by the Bayesian inferring framework, the FA-ET-PHD filter can effectively avoid the deterioration of the performance in multi-target tracking caused by the nonlinear change of the distribution of scatters. Simulation results show that the proposed method can improve the accuracy of the multi-target state estimation as well as the robustness.