Radar Signal Recognition Based on MSST and Dual Channel Feature Extraction

Aiming at the problems of the traditional low probability of intercept (LPI) radar signal recognition algorithm, which is difficult to extract effective features and low recognition accuracy under low signal-to-noise ratio (SNR), a radar signal recognition method based on multi-synchrosqueezing transform (MSST) and two channel feature extraction was proposed. First, MSST is used to obtain signal time frequency images, and perform a series of pre-processing operations. Then, the image features are extracted through the gray co-incidence matrix (GLCM) and the local binary variance mode (LBPV), and the feature dimension is reduced by the principal component analysis (PCA). Finally, the dimensionality reduction features are sent to the support vector machine (SVM) to complete the classification and recognition of radar signals. According to the simulation analysis results, the recognition accuracy of the two-channel feature extraction algorithm is 15.67% higher than that of the two single channels on average under a low SNR and significantly higher than that of the symmetric Holder coefficient method. Moreover, when the SNR is as low as -2dB, the overall recognition rate of the algorithm is still 96.44%. This algorithm not only has low computational complexity but also has high recognition accuracy, especially compared with single-channel feature extraction and symmetric Holder coefficient methods at low SNR, the recognition accuracy has significant advantages.