Multiple Targets Localization Behind L-Shaped Corner via UWB Radar
Songlin Li,Shisheng Guo,Jiahui Chen,Xiaqing Yang,Shihao Fan,Chao Jia,Guolong Cui,Haining Yang
DOI: https://doi.org/10.1109/tvt.2021.3068266
IF: 6.8
2021-04-01
IEEE Transactions on Vehicular Technology
Abstract:This paper deals with the multiple targets localization problem via multi-channel ultra-wideband (UWB) imaging radar non-line-of-sight (NLOS) signal processing. A novel matching-based radar imaging algorithm is proposed to obtain the positions of multiple targets in the L-shaped corner scenario with complex multipath ghost signals. Firstly, a multipath propagation model for the multiple targets scenario is established. Then the positions of the actual multipath ghosts are extracted from the radar image, and the candidate targets corresponding to these multipath ghosts are derived. Secondly, the ellipse-cross-localization method is proposed to obtain the positions of the candidate multipath ghosts, followed by two defined matching factors to measure the similarity between actual and candidate multipath ghosts. According to the similarity, decision rules are designed to determine the actual targets. Compared with the localization algorithm based on one-dimensional range profile, the proposed algorithm can effectively cope with the cases of multiple targets, even in the cases of rough walls and noise. Finally, simulations and experimental data are used to validate the effectiveness of the proposed algorithm.
telecommunications,engineering, electrical & electronic,transportation science & technology