基于VMD联合RCMDE的特定辐射源识别方法

A Specific Emitter Identification Method Based on VMD and RCMDE

针对常用于特定辐射源识别(Specific Emitter Identification,SEI)的典型一维特征常常引发识别性能下滑问题,高维度特征维度较大、与一般分类器结合使用时计算效率较低的问题,提出了一种基于变分模态分解(Variational Mode Decomposition,VMD)和精细复合多尺度散布熵(Refined Composite Multi-scale Dispersion Entropy,RCMDE)的SEI方法,利用VMD和RCMDE获取原始辐射源信号不同频率分量的多尺度时间复杂度特征,选择支持向量机(Support Vector Machine,SVM)完成分类识别。仿真结果表明,莱斯信道下,在-5~15 dB的信噪比(Signal-to-Noise,SNR)范围内,所提方法对3个不同辐射源个体的识别准确率达到了99.2367%,相比于其他方法有显著的性能提升。

To solve the problems that the typical one-dimensional features commonly used in Specific Emitter Identification(SEI)often lead to the decline of recognition performance,large dimensions of high-dimensional features and low computational efficiency when combined with general classifiers,an SEI method based on Variational Mode Decomposition(VMD)and Refined Composite Multi-scale Dispersion Entropy(RCMDE)is proposed.VMD and RCMDE are used to obtain the multi-scale time complexity characteristics of different frequency components of the original emitter signal.Finally,the SVM is selected to complete the classification.The simulation results show that in the range of Signal-to-Noise Ratio(SNR)from-5 dB to 15 dB in Riacian channel,the recognition accuracy of the method for three different radiation sources is 99.2367%.Compared with other methods,the performance is significantly improved.