A Novel Adversarial Learning Framework for Passive Bistatic Radar Signal Enhancement

Passive Bistatic Radar (PBR) has significant civilian and military applications due to its ability detect low-altitude targets. However, the uncontrollable characteristics of transmitter often lead subpar target detection performance, primarily a low signal-to-noise ratio (SNR). Coherent accumulation typically limited improve SNR in presence strong noise clutter. In this paper, we propose an adversarial learning-based radar signal enhancement method, called generative network (RSEGAN), overcome challenge. On one hand, encoder-decoder structure is designed map noisy signals clean ones without intervention training stage. other hybrid loss constrained by L1 regularization, L2 gradient penalty proposed ensure effective RSEGAN. Experimental results demonstrate that RSEGAN can reliably remove from information, providing gain higher than 5 dB for basic coherent integration method even under conditions.