A Radar Signal Modulation Recognition Method Based on SSAGAN

The actual battlefield where the radar signal is located is a non-cooperative confrontation environment, and the radar signals intercepted by reconnaissance aircraft have known and unknown modulation types. The recognition of different modulation types of radar signals plays a vital role in the actual battlefield environment. In order to recognize radar signals with known modulation types and unknown modulation types, this paper designs a Spatial Spectrum Attention Generative Adversarial Network (SSAGAN) model. SSAGAN uses encoder structure to replace the generation model in Generative Adversarial Network (GAN). The Shuffle Attention (SA) mechanism is introduced to improve the effectiveness of image information acquisition. Simulation results show that the model can effectively improve the recognition accuracy of different modulation types of radar signals.