基于改进OPTICS聚类的雷达信号预分选方法  

Radar Signal Pre-sorting Method Based on Improved OPTICS Clustering

作　　者：吴连慧[1] 周秀珍[1] 宋新超[1] WU Lian-hui;ZHOU Xiu-zhen;SONG Xin-chao(The 723 Institute of CSIC,Yangzhou 225101,China)

机构地区：[1]中国船舶重工集团公司第七二三研究所,江苏扬州225101

出　　处：《舰船电子对抗》2018年第6期95-99,共5页Shipboard Electronic Countermeasure

摘　　要：针对DBSCAN算法不能分选密度分布不均雷达信号的缺陷,提出了一种基于改进OPTICS聚类的雷达信号预分选方法。利用OPTICS算法的核心距离和可达距离来实现对不同密度分布的雷达信号预分选。将雷达信号进行网格单元划分,并采用两级处理方式,第一级处理高密度信号,第二级累积并处理低密度信号。实验结果表明,与传统的DBSCAN算法相比,改进OPTICS算法适用于不同密度分布的雷达信号,能够提高复杂体制雷达信号分选的正确率,同时降低了算法运行时间。Aiming at the defect that the radar signal of uneven density distribution can not be sorted by DBSCAN algorithm,this paper puts forward a radar signal pre-sorting method based on improved OPTICS clustering,uses the core range and reachable range of OPTICS algorithm to realize the radar signal pre-sorting of different density distribution,divides radar signal into grid cells,adopts 2 stages processing methods:in the first stage,the high density signals are processed;while in the second stage,the low density signals are accumulated and processed.The experiment result shows that:compared with traditional DBSCAN algorithm,the improved OPTICS algorithm is fit for radar signal of different density distribution,and can raise the accuracy of radar signal sorting with complex system,as well as reduce the operating time of algorithm.

关 键 词：雷达信号分选 预分选 OPTICS算法 网格单元 两级处理 

分 类 号：TN971.1[电子电信—信号与信息处理]