基于多域特征提取和RVM的有源压制干扰识别方法  

Identification Method of Active Blanket Jamming Based on Multi-domain Feature Extraction and RVM

作　　者：蔡潇 李大超 翁永祥 王明君 CAI Xiao;LI Da-chao;WENG Yong-xiang;WANG Ming-jun(The 10th Military Representative Office of Navy Equipment Department in Shanghai,Shanghai 201800,China;The 51st Research Institute of CETC,Shanghai 201802,China)

机构地区：[1]海军驻上海地区第十军事代表室,上海201800 [2]中国电子科技集团公司第五十一研究所,上海201802

出　　处：《舰船电子对抗》2021年第2期23-27,47,共6页Shipboard Electronic Countermeasure

摘　　要：针对有源压制干扰信号的自动识别问题,提出一种基于多域特征提取和相关向量机(RVM)分类器的识别方法。首先对噪声干扰,密集假目标干扰,间歇采样转发干扰和组合干扰4种压制干扰信号进行建模和分析,然后分别提取时域特征、频域特征和变换域特征构成多域特征向量,最后采用RVM分类器进行特征选择和分类识别,自动确定最优识别特征的同时提升算法稳健性。基于仿真数据的实验结果表明,相对于传统单一维度特征,所提多域特征可以获得更优的识别性能,并且在低JNR条件下具有更高的鲁棒性。Aiming at the problem of automatic recognition of active blanket jamming signal,a recognition method based on multi-domain feature extraction and relevance vector machine(RVM)classifier is proposed.Firstly,the modeling and analysis of noise jamming,dense false target jamming,intermittent sampling repeater jamming and combined jamming are performed.Then the time domain feature,frequency domain feature and transform domain feature are extracted to form multi-domain feature vector.Finally,RVM classifier is used for feature selection and classification recognition,which automatically determines the optimal recognition feature and improves the robustness of the algorithm.The experimental results based on simulation data show that the proposed method based on multi-domain feature can achieve better recognition performance,and have higher robustness under low jamming-to-noise ratio(JNR)conditions than that based on traditional single dimension feature.

关 键 词：雷达对抗 压制干扰 模式识别 特征提取 相关向量机 

分 类 号：TN972.1[电子电信—信号与信息处理]