Radar HRRP Estimation Based on Deep Unfolding Networks under Interrupted Sampling Repeater Jamming

Radar high-resolution range profile (HRRP) estimation under interrupted sampling repeater jamming (ISRJ) is a challenging task in complicated environments. In this paper, we modify two deep unfolding networks (DUNs), i.e., analytical learned iterative soft thresholding algorithm (ALISTA) and analytical learning proximal operator method (ALePOM), to work under various situations. We first detect and remove the interfered components in the received signal and then employ the modified DUNs to estimate the HRRP of target. Specifically, by taking pre-calculated matrices as input data, these two modified DUNs overcome the challenges of parameter tuning for traditional iterative sparse recovery methods and the dependence on specific dictionaries for existing DUNs, enabling their applications to various situations with changing radar parameters and/or jamming parameters. Furthermore, we outline the dataset construction and training schemes to improve the performance and reduce the training difficulty of the proposed methods. Experiment results demonstrate that the proposed methods exhibit superior effectiveness and advantages in radar HRRP estimation under ISRJ.