基于自适应仿射传播聚类的雷达信号分选方法  

A method of radar signal sorting based on adaptive affinity propagation clustering

作　　者：陈昌云[1] 郑子扬[1] 陈永游[1] 

机构地区：[1]中国航天科工集团8511研究所,江苏南京210007

出　　处：《航天电子对抗》2011年第3期20-22,35,共4页Aerospace Electronic Warfare

摘　　要：复杂体制的雷达不断应用于实际,使得基于到达时间单参数的雷达信号分选方法遇到了技术瓶颈。然而,基于多参数的信号分选方法,如K-均值聚类和模糊函数C均值聚类,前者对初始类中心敏感、需预知聚类数,后者对类隶属度的选取、相似度矩阵的建立存在不确定性,使得这些方法在自适应雷达信号分选中受到了限制。自适应仿射传播聚类不仅可以克服上述方法的不足,而且可以自适应地给出最优聚类数,因此,可将自适应仿射传播聚类应用于雷达信号分选中。仿真实验表明了该方法的有效性。The use of radars with complicated systems makes the signal sorting method based on the time of arrival encountering a bottleneck. However, the signal sorting methods based on multi-parameter such as K- means and fuzzy C means （FCM） clustering are also limited in adaptive radar signal sorting. The faults are that the K-means is sensitive to the initial cluster center and requires predicting the number of clusters, and how to select the class membership and establishment of similarity matrices in FCM is uncertain. Adaptive affinity propagation clustering （AdAP） method can not only overcome these faults, but also gives the optimal number of adaptive clustering. Therefore, the method is applied to radar signal sorting. The simulation results prove the validity of the proposed method.

关 键 词：信号分选 自适应 仿射传播 聚类 

分 类 号：TN957.51[电子电信—信号与信息处理]