Periodic ISRJ suppression with convolution generative network in delay-Doppler domain

Based on digital radio frequency memory (DRFM), interrupted sampling repeater jamming (ISRJ) poses a serious threat to radar systems by generating false targets after matched filtering. The suppression of jamming has become an intriguing topic in recent years. In this paper, the distribution feature of false targets in the delay-Doppler domain is derived by analyzing the target operator and performing Fourier series expansion on the periodic sampling pulse of ISRJ. The solution for suppressing ISRJ is then transformed into a matrix fitting problem, which may require two-dimensional convolution and feature extraction. With the aid of convolutional generative network (CGN), the delay-Doppler matrix of false targets can be reconstructed and cancelled with the original jammed matrix. The ISRJ-suppressed pulse compression output can finally be extracted from the jamming-cancelled matrix. Simulation experiments on three kinds of common radar waveforms are conducted to confirm the effectiveness and universal applicability of the proposed method.