基于经验模态分解的通信辐射源分形特征提取算法

摘要

In view of the problem that the accuracy of the individual identification method of communication transmitter is low,and poor in robust,a feature extraction method based on the empirical mode decomposition(EMD)model was proposed in this paper.In this algorithm,fractal features of intrinsic mode function from the time domain and frequency domain was extracted,which combined with the Hilbert marginal spectrum fractal feature and symmetry coefficient to get feature vector.Support vector machine(SVM)was employed to identify transmitter individuals.The experimental results of the classification of the 10 parts of the proposed radio station showed that the algorithm could obtain better classification results without any prior information and had a certain robustness.%针对当前通信辐射源个体识别方法精度不高，鲁棒性不强等问题，提出了基于经验模态分解(EMD)模型的通信辐射源特征提取算法。该算法通过提取包括本征模函数(IMF)时域和频域范围内的分形特征结合Hilbert 边缘谱上的分形特征与谱对称系数组成特征向量，并使用支持向量机(SVM)得到分类结果。10部建伍电台的分类实验结果表明：该算法在不需要先验信息的前提下，可以得到较好的分类效果，并且具有一定的鲁棒性。