基于距离特征的雷达辐射源信号识别方法  

Radar Emitter Signal Identification Via Distance Features

作　　者：黄颖坤 金炜东[1,2] 颜康 朱劼昊 Huang Yingkun;Jin Weidong;Yan Kang;Zhu Jiehao(College of Electrical Engineering,Southwest Jiao tong University,Chengdu 610031,China;Science and Technology on Electronic Information Control Laboratory,Chengdu 610031,China)

机构地区：[1]西南交通大学电气工程学院,四川成都610031 [2]电子信息控制重点实验室,四川成都610031

出　　处：《系统仿真学报》2021年第12期2959-2966,共8页Journal of System Simulation

基　　金：电子信息控制重点实验室开放基金(6142105190312)。

摘　　要：针对传统的雷达辐射源信号识别方法在低信噪比环境下的正确率较低,且通常只适用几种特定的雷达信号的问题,提出一种基于距离特征的辐射源信号识别方法。使用k-means算法提取若干个聚类中心,分别计算雷达信号脉冲与聚类中心之间的DTW (Dynamic Time Warping)度量值,联合这些度量值作为k邻近算法的输入进行识别。仿真结果表明,在信噪比为3d B时,所提方法对6类雷达信号的识别率达到91%。与基于小波脊频级联特征的方法相比,所提方法也表现出更好的识别效果。Aiming at the problem that traditional recognition methods of radar emitter signal have low accuracy in low signal to noise ratio(SNR) environment, and are usually suitable for only several specific radar signals, an identification approach of radar signal based on distance features is proposed. Several cluster centers are extracted via the k-means algorithm, and the Dynamic Time Warping(DTW) values between the radar signal and the cluster center are calculated respectively, which are combined as the input features of k-Nearest Neighbor(k-NN) algorithm. The simulation results show that when the SNR is 3 d B, the identification rate of the 6 classes of radar signals is 91%. Compared to the method based on wavelet ridge-frequency cascade-feature, the proposed method also shows better recognition performance.

关 键 词：雷达辐射源信号识别 聚类中心 DTW(Dynamic Time Warping)度量方法 k邻近算法 距离特征 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]