基于时频域细微特征的辐射源个体识别  

Specific Emitter Identification Based on Time-Frequency Domain Characteristic

作　　者：王欢欢[1] 张涛[1] 孟凡玉[1] WANG Huanhuan,ZHANG Tao(Information Engineering University, Zhengzhou 450001 , China)

机构地区：[1]信息工程大学,河南郑州450001

出　　处：《信息工程大学学报》2018年第1期23-29,共7页Journal of Information Engineering University

基　　金：科研基金资助项目

摘　　要：在基于Huang提出的希尔伯特黄变换(Hilbert-Huang transform,HHT)分析辐射源个体信号时频特征的基础上,针对HHT存在的端点效应问题,提出一种对瞬时频率进行估计的改进HHT算法(improved HHT,IHHT)。仿真实验验证了IHHT算法能够得到信号比较精确的瞬时特征。用IHHT算法得到信号的Hilbert时频谱,并提取实际辐射源个体通信信号的几种时频细微特征,采用有监督学习分类器支持向量机(support vector machine,SVM)、交叉验证搜索最佳参数方法进行分类识别,与传统方法相比,IHHT算法能够比较精确地描述信号的时频特征,平均识别率提高到90%以上。Hilbert-Huang transform has the weakness of endpoint ettect. To deal with this problem, this paper proposes an improved HHT（IHHT） to estimate the instantaneous frequency. Then it is proved that the IHHT method can get accurate transient signal characteristics by simulation experi- ments, and the endpoint effect has been effectively suppressed. Finally this paper uses IHHT algo- rithm to extract time-frequency characteristics, uses supervised learning SVM classifier for identif- ying, and compares the recognition results with the traditional HHT methods. Recognition results show IHHT can describe the time-frequency characteristics exactly, and the average recognition rate can reach no less than 90%.

关 键 词：辐射源识别 希尔伯特黄变换 瞬时特征 支持向量机 

分 类 号：TN911.6[电子电信—通信与信息系统]