基于自适应滑动窗均值偏移算法的雷达信号分选  

Radar signal sorting based on an auto-mean-shift algorithm

作　　者：郭立民[1] 陈昊翔 于飒宁 GUO Limin;CHEN Haoxiang;YU Saning(College of Information and Communication Engineering,Harbin Engineering University,Harbin 150001,China)

机构地区：[1]哈尔滨工程大学信息与通信工程学院,黑龙江哈尔滨150001

出　　处：《哈尔滨工程大学学报》2024年第11期2266-2273,共8页Journal of Harbin Engineering University

摘　　要：为了应对复杂体制雷达信号分选的要求,提升信号分选的速度和准确性,解决传统信号分选方法的弊端,本文将均值偏移聚类算法引入雷达信号分选,通过改进算法引入滑动窗的自适应更新摆脱了传统分选方法对预设参数的依赖。实验结果表明:自适应滑动窗均值偏移算法在分选准确性和速度上综合表现优于用于对比的传统分选方法,对不同密度的雷达数据集和在脉冲丢失的情况下都可实现较好的分选效果,对捷变频雷达的分选效果良好,并可推知对其他类似体制的雷达均有良好分选效果。本文研究成果可应用于大量数据的多体制雷达的分选优化。To address the complex requirements of institutional radar signal sorting,improve the speed and accuracy of signal sorting,and overcome the drawbacks of traditional methods,this paper introduces the mean-shift clustering algorithm into radar signal sorting.Improving the algorithm with adaptive updating of the sliding window eliminates the dependence on preset parameters typical of traditional sorting methods.Experimental results show that the auto-mean-shift algorithm outperforms the traditional sorting method in terms of sorting accuracy and speed.It achieves better sorting performance for radar data sets with different densities and under conditions of pulse loss.Additionally,it performs well with agile frequency radars,suggesting its effectiveness for other similar radar systems.The research results of this paper suggest that this method can be applied to optimize the sorting optimization of multi-regime radar systems handling large volumes of data.

关 键 词：雷达信号分选 脉冲描述字 机器学习 聚类算法 均值偏移聚类 捷变频雷达 K-MEANS算法 DBSCAN算法 

分 类 号：TN971.1[电子电信—信号与信息处理]