基于典型相关分析的雷达信号脉内特征识别  

Recognition of Intra-Pulse Feature of Radar Signals Based on Canonical Correlation Analysis

作　　者：李咏晋 赵拥军[2] 赵闯[2] LI Yongjin1 , ZHAO Yongjun2, ZHAO Chuang2(1 Luoyang Electronic Equipment Test Center of China, Luoyang 471003, China 2. Information Engineering University, Zhengzhou 450001, Chin)

机构地区：[1]洛阳电子装备试验中心,河南洛阳471003 [2]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47-51,共5页Journal of Information Engineering University

摘　　要：雷达信号的模糊函数是雷达辐射源识别中一种重要的时频特征。为解决低信噪比下辐射源识别率较低的问题,提出基于雷达信号模糊函数原点切片特征的识别算法。结合对分数阶傅立叶变换的分析,采用分数阶傅立叶变换计算出模糊函数的原点切片,将切片构建成互补特征集对后,利用典型相关分析与核典型相关分析算法,消除模糊函数特征的冗余,改善分选的准确率与抗噪声性能。实验结果表明上述方法在较低信噪比下能够有效提高识别性能。Ambiguity function （AF） is an important time-frequency teature of radar emitters recog- nition. A recognition algorithm based on Origin-cuts of AF is proposed for the low radar emitters rec- ognition rate under low SNRs. Firstly, the Origin-cuts of AF is calculated by fractional Fourier trans- form （FRFT） with the analysis of FRFT. Then, two different pairs of teature vectors with comple- mentary information can be constructed. Canonical correlation analysis （CCA） or kernel canonical correlation analysis （KCCA） is used to remove the redundancy of AF and to improve the sorting pre- cision and noise immunity. Experimental results show the algorithms improve the recognition per- formance considerably under low SNRs.

关 键 词：雷达辐射源识别 模糊函数 分数阶傅立叶变换 典型相关分析 

分 类 号：TN957.51[电子电信—信号与信息处理]