Research on classification and recognition of electronic warfare signals based on neural network

Signal classification and recognition is one of the important technologies in electronic warfare, the traditional signal classification and recognition has limitations, it can only be carried out under simple conditions, and artificial intelligence (AI) can break the limitations of signal identification. In this paper, the fusion of multi-scale time-frequency domain CNN, Transformer and LSTM-CNN is applied to the identification and recognition of multi-wavelength and complex encoded signals. Self-attention mechanism, reverse training and lightweight network are introduced to improve the classification effect of high signal-to-noise ratio, time-varying signals and embedded systems. The artificial intelligence signal processing method based on AI breaks through the limitations of traditional algorithms and improves the timeliness and accuracy of signal recognition.