Time-frequency image and high-order spectrum characteristics based radar signal recognition

Abstract

Aiming at improving the accuracy of radar signal recognition under a low signal-to-noise ratio, a radar signal recognition algorithm based both on time-frequency image and high-order spectrum feature was proposed.Firstly, the time-frequency image was obtained by Choi-Williams distribution (CWD) transform, based on which the time-frequency image was preprocessed and the texture features were extracted by gray level co-occurrence matrix (GLCM) in sequence.Meanwhile, the symmetrical holder coefficient was used to extract the high-order spectral features of the signal.Then, the texture features and high-order spectrum features were form a new set of joint feature vectors.Finally, with the proposed feature vector the classification and recognition of radar signals were implemented by a support vector machine.The algorithm was verified on the data set with eight typical radar signals.Experimental results show that the recognition accuracy of different radar signals can achieve higher than 90% when the signal-to-noise ratio is -8 dB.

Keywords

• radar signal recognition;high order spectrum;Choi-Williams time frequency distribution;support vector machine