太赫兹科学与电子信息学报2025，Vol.23Issue(11)：1130-1140,11.DOI:10.11805/TKYDA2024226
基于杂波稀疏特征提取的外源雷达目标探测
Sparse clutter extraction-based target detection for passive radar
摘要
Abstract
A target detection scheme is proposed to address the problem of weak target echo compared to clutter in Orthogonal Frequency Division Multiplexing(OFDM)-based passive radar.This involves first estimating the sparse features of clutter in the Range-Doppler(RD)domain using deep networks and then subtracting them from the original RD data to extract weak target information.Firstly,a novel sparse model in the RD domain is proposed,enabling the reconstruction network to directly output sparse component coefficients and meet the condition of similarity between the reconstructed object and the true value,thus eliminating the need to construct a dictionary matrix that fully matches the time-varying transmission waveform.Subsequently,a physically interpretable deep unfolding network,Long Range(LR)-Iterative Shrinkage Thresholding Algorithm(ISTA)-Net,is established,basing on ISTA.This network can realize automatically learning of parameters such as dictionary matrices,step sizes,and thresholds,and it has only 7%of the parameter quantity compared to the original Learnable ISTA(LISTA)deep unfolding network.Additionally,a new threshold function Lsoft is introduced to enable LR-ISTA-Net to better learn sparse features in clutter data.Finally,simulations and actual measurements are employed to validate the effectiveness and superiority of LR-ISTA-Net to extract target information from passive radar.关键词
正交频分复用波形外源雷达/目标检测/稀疏模型/迭代收缩阈值算法/杂波特征提取Key words
Orthogonal Frequency Division Multiplexing(OFDM)-based passive radar/target detection/sparse model/Iterative Shrinkage Thresholding Algorithm(ISTA)/clutter feature extraction分类
信息技术与安全科学引用本文复制引用
赵志欣,陈方越,曹玉龙,李周章,万跃辉..基于杂波稀疏特征提取的外源雷达目标探测[J].太赫兹科学与电子信息学报,2025,23(11):1130-1140,11.基金项目
国家自然科学基金资助项目(62261036) （62261036）
江西省自然科学基金资助项目(20242BAB23010) （20242BAB23010）
