Underwater acoustic communication signal modulation identification method and system based on time-frequency diagram

Abstract

The invention provides an underwater acoustic communication signal modulation identification method and system based on a time-frequency diagram. The method comprises the following steps: preprocessing an acquired underwater acoustic communication signal; generating a time-frequency image from the preprocessed underwater acoustic communication signal by using a short-time Fourier transform method; inputting the time-frequency image into a trained generative adversarial network denoising model based on an attention mechanism, and outputting a denoised time-frequency image; and inputting the denoised time-frequency image into a trained underwater acoustic communication signal feature extraction and recognition model based on the convolutional neural network to obtain a signal recognition result. According to the method, the problem that modulation recognition of the underwater acoustic communication signals is difficult in a noise interference environment is effectively solved, the attention mechanism guides the model to pay attention to the key feature area of the time-frequency graph, and the recognition precision is improved.

Images