Radar emitter signal classification based on mutual information and fuzzy support vector machines

In this paper, a novel method based on mutual information and fuzzy support vector machines for recognizing radar emitter signals is introduced. The radar signal waveforms are the linear frequency modulation (LFM), frequency-coded signals, BPSK and QPSK. The wavelet ridges and higher-order statistics are used to extract signal features. Mutual information measures is used to reduce the redundant components from the feature vectors set. Then these discriminative and low dimensional features achieved are fed to a fuzzy support vector machine classifier for multi-class patter recognition. In simulation, the classifier attains over 78% overall average correct classification rate. Experimental results show that the proposed methodology is efficient for different complex radar signals detection and classification.