基于小波脊和FSVM的雷达辐射源识别

Radar emitter recognition based on wavelet ridge and FSVM

有效的特征提取和信号特征选择是解决复杂体制雷达辐射源信号分选难题的重要手段。利用小波脊和高阶谱分析方法提取雷达辐射源信号的瞬时频率、瞬时相位和幅度以及高阶累积量等特征向量。通过基于互信息的贪婪算法进行特征选择,得到具有低维数、可识别性的辐射源特征。为解决多分类问题中的不可分情况,引入基于模糊C均值聚类的模糊支持向量机进行雷达辐射源分类识别实验。实验表明,该方法对多种复杂辐射源信号具有较好的识别效果。

Effective feature extraction and selection are dominant measures to solve the issues of radar emitter signal sorting and classification.Wavelet ridge and high order spectrum analysis are used to extract the features such as the instantaneous frequency,instantaneous phase and amplitude.The feature selection algorithm based on mutual information is provided.Then these obtained discriminative and low dimensional features are fed to a support vector machine classifier based on FCM clustering for multi-class pattern recognition.Experiment results show that the proposed method is efficient for the detection and classification of various complex radar emitter signals.