Feature Extraction and Identification of Specific Radiation Sources Based on Axial Integral Bispectrum and Deep Residual Shrinkage Network

In the existing research on the identification of specific radiation sources, some noise inevitably occurs when classifying samples, which affects the extraction of Radio frequency fingerprint (RFF) with unique native properties, thus reducing the classification accuracy. In this paper, a feature extraction and identification method for specific radiation sources based on the integration of axial integral bispectrum features and deep residual shrinkage networks(DRSN) is proposed. First, the axial integral bispectrum is used to extract the signal features, and then the signal is denoised by soft thresholding in the deep residual shrinkage network, and the threshold is automatically set according to the situation of each sample using the attention mechanism. The experimental results show that the new method can effectively improve the classification accuracy under low signal-to-noise ratio, and in the case of low signal-to-noise ratio, the maximum distance of 62ft can achieve 98.5% classification and recognition accuracy.