A Target Detection and Tracking Method for Multiple Radar Systems.
Bo Yan,Enrico Paolini,Luping Xu,Hongmin Lu
DOI: https://doi.org/10.1109/tgrs.2022.3183387
IF: 8.2
2022-01-01
IEEE Transactions on Geoscience and Remote Sensing
Abstract:Multiple radar systems represent an attractive option for target tracking because they can significantly enlarge the area coverage and improve both the probability of trajectory detection and the localization accuracy. The presence of multiple extended targets or weak targets is a challenge for multiple radar systems. Moreover, their performance may be severely deteriorated by regions characterized by a high clutter density. In this article, an algorithm for detection and tracking of multiple targets, extended or weak, based on measurements provided by multiple radars in an environment with heavily cluttered regions, is proposed. The proposed method features three stages. In the first stage, past measurements are exploited to build a spatiotemporal clutter map in each radar; a weight is then assigned to each measurement to assess its significance. In the second stage, a track-before-detect algorithm, based on a weighted 3-D Hough transform, is applied to obtain target tracklets. In the third stage, a low-complexity tracklet association method, exploiting a lion reproduction model, is applied to associate tracklets of the same target. Three experiments are presented to illustrate the effectiveness of the proposed approach. The first experiment is based on synthetic data, the second one is based on actual data from a radar network with two homogeneous air surveillance radars, and the third one is based on actual data from a radar network with four different marine surveillance radars. The results reveal that the proposed method can outperform competing approaches.