Target Detection and Interference Cancellation in Passive Radar Sensors via Group-Sparse Regression

Passive radar sensors utilize communication signals that are not specifically designed for the purpose of target detection and localization. As a consequence, the generated range-Doppler map exhibits numerous sidelobes alongside the primary peaks. Thus, the echoes of strong targets and clutter from these sidelobes can easily mask targets with weaker signals. Several existing algorithms attempt to suppress these echoes from the received signal before detecting targets. In this article, we observe that in some situations these algorithms generate artifacts that are sometimes falsely detected as targets but also sometimes mask weaker targets and reduce the detection probability. As a remedy, we utilize the sparsity of received signals in the range-Doppler and employ a stepwise regression algorithm to extract all scatterers one by one. Moreover in this algorithm, we consider off-grid scatterers (OGSs). To archive an acceptable performance, we must consider a considerable number of signal samples and a large number of possible range-Doppler combinations which makes the algorithm computationally expensive. Therefore, we exploit the structure of the involving matrices and propose a very efficient algorithm that simultaneously detects targets, clutter, and direct-path. Our simulations demonstrate the superior performance of the proposed algorithm in target detection and interference cancellation (IC).