基于主成分分析的雷达行为状态聚类分析方法  

Radar Behavior State Cluster Analysis Based on Principal Component Analysis

作　　者：方旖 毕大平[1] 潘继飞 陈秋菊 FANG Yi;BI Daping;PAN Jifei;CHEN Qiuju(Institute of Electronic Countermeasure,National University of Defence Technology,Hefei 230037,China;Laboratory of Electronic Countermeasure Information Processing,National University of Defence Technology,Hefei 230037,China)

机构地区：[1]国防科技大学电子对抗学院预警对抗系,安徽合肥230037 [2]电子对抗信息处理实验室,安徽合肥230037

出　　处：《探测与控制学报》2020年第2期112-118,共7页Journal of Detection & Control

基　　金：国家自然科学基金项目资助(61671453)。

摘　　要：针对在实际作战中雷达对抗侦察设备对未知雷达进行工作状态识别时,难以获取和利用充足的训练样本的问题,提出了基于主成分分析的雷达行为状态聚类分析方法。该方法通过分析雷达行为状态类型与雷达信号脉冲参数间的本质关联,以获得可用于分类的特征。首先对机载侧视雷达远区搜索、近区搜索及跟踪等不同工作状态下的雷达信号特征进行深入分析,选择合适的雷达脉冲序列组成雷达样本库。然后,对测试信号特征进行预处理,结合主成分分析方法提取合适的特征参数,并对C-均值聚类算法进行了改进,完成了样本数据的分类,该方法能够在小样本数据条件下进行分类。仿真结果表明,该方法可以用于雷达行为状态的分类并且不易受初始数据及随机聚类中心的影响。Aiming at the problem that it was difficult to acquire and utilize sufficient training samples when radar countermeasure reconnaissance equipment was used to identify the working state of unknown radar in actual combat,a method of radar behavior state cluster analysis based on principal component analysis was proposed.By analyzing the essential correlation between radar behavior state types and radar signal pulse parameters,this method obtained the characteristics that could be used for classification.Firstly,the signal characteristics of an airborne side-looking radar under different working conditions,such as far region search,near region search and tracking,were analyzed in depth.Then,the characteristics of the test signal were preprocessed,and the appropriate characteristic parameters were extracted by combining the principal component analysis method,and the c-mean clustering algorithm was improved to complete the classification of sample data.Simulation results showed that this method could be used to classify radar behavior and was not easily affected by initial data and random clustering center.

关 键 词：认知对抗雷达 雷达行为 C-均值算法 主成分分析 聚类识别 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构] TN957.51[自动化与计算机技术—计算机科学与技术]