Anti-Jamming Method of Near-Field Underwater Acoustic Detection Based on WGAN

This paper analyzes the issue of artificial interference encountered in underwater near-field detection. We briefly examines three common types of artificial jamming signal and their mechanisms. Taking inspiration from the application of Generative Adversarial Networks in speech signal enhancement, this study employs Wasserstein GAN and integrates the characteristics of detection signals. L2 loss is added to the generator's loss function in WGAN to enhance training stability. Simulation analysis demonstrates that the trained WGAN generator effectively combats the three types of artificial jamming.