用于雷达信号分选的K中位最近邻聚类算法  

K-median nearest neighbor clustering algorithm for radar signal sorting

作　　者：伍佳钰 甄佳奇[1] WU Jiayu;ZHEN Jiaqi(College of Electronic Engineering,Heilongjiang University,Harbin 150080,China)

机构地区：[1]黑龙江大学电子工程学院,哈尔滨150080

出　　处：《黑龙江大学自然科学学报》2024年第4期496-504,共9页Journal of Natural Science of Heilongjiang University

基　　金：黑龙江省自然科学基金资助项目(LH2022F043)。

摘　　要：在处理雷达信号时,基于密度的空间聚类(Density-based spatial clustering of applications with noise,DBSCAN)分选算法依赖于参数或阈值的选取,影响分选的准确率。为此提出了一种改进的雷达信号脉冲分选算法,在DBSCAN聚类基础上结合了K中位最近邻(K-median nearest neighbor,KMNN)算法,通过引入自衰减系数并设置阈值上限对参数值列表进行二次处理,可以自适应根据聚类结果与不同参数时的K值之间的关系确定最优的邻域半径和最少点个数,提高了分选的正确率。通过仿真实验验证了算法利用雷达脉冲描述字特征进行自适应分选的有效性。When processing radar signals,the Density-based spatial clustering of applications with noise(DBSCAN)algorithm depends on the selection of parameters or thresholds,which affects the accuracy of the sorting.To solve this problem,the improved radar signal pulse sorting algorithm is proposed.On the basis of DBSCAN clustering,K-median nearest neighbor(KMNN)algorithm is combined.The self-attenuation coefficient is introduced and the upper threshold is set for secondary processing of the parameter value list.According to the relationship between the clustering results and the K values of different parameters,the optimal neighborhood radius and the minimum number of points are determined adaptively,and the accuracy of sorting is improved.The effectiveness of the algorithm based on the radar pulse descriptor character is verified by simulation experiments.

关 键 词：雷达信号分选 聚类 DBSCAN K中位最近邻算法 

分 类 号：TN971[电子电信—信号与信息处理]