| 摘要: |
| 为了在复杂多变的电子战场景下对密集重叠的雷达脉冲信号进行快速准确的分选，稀释脉冲流，解决现有基于密度的空间聚类算法(Density-based Spatial Clustering of Applications with Noise,DBSCAN)在分选时易受干扰点影响、聚类参数需要人为设置、算法复杂度高的问题，提出了一种面向雷达信号预分选的粒子群快速密度聚类算法(Particle Swarm Fast Density Clustering Algorithm,PSK-DBSCAN)。该算法首先引入数据场理论剔除雷达脉冲信号里的干扰点，提升了分选准确度；其次，引入粒子群算法并设计了基于轮廓系数的适应度函数，自适应地获得最优聚类参数；最后，使用K-D(K-Dimensional)树降低DBSCAN的算法复杂度，减少分选时间。经实验验证，算法可以对复杂交错的雷达脉冲信号实现快速聚类分选，正确率达到98.9%，性能稳定。 |
| 关键词: 雷达信号分选 数据场 粒子群算法 K-D树 密度聚类 |
| DOI：10.20079/j.issn.1001-893x.240618001 |
|
| 基金项目:中国科学院国防重点实验室基金 |
|
| A Particle Swarm Fast Density Clustering Algorithm for Radar Signal Pre-sorting |
| LU Xinyu,HUANG Yonghui,CUI Tianshu,ZHU Yan,HAN Jiabao |
| (1.Key Laboratory of Electronics and Information Technology for Space System，National Space Science Center，Chinese Academy of Science，Beijing 101499，China;2.University of Chinese Academy of Sciences，Beijing 100049，China;3.China Academy of Aerospace Science and Innovation，Beijing 100048，China) |
| Abstract: |
| To efficiently sort dense and overlapping radar pulse signals in dynamic electronic warfare environments,a particle swarm fast density clustering algorithm(PSK-DBSCAN) is proposed.This algorithm addresses the limitations of existing density-based spatial clustering of applications with noise(DBSCAN) radar signal sorting methods,such as susceptibility to interference,reliance on manual parameter tuning and high computational complexity.The algorithm first applies data field theory to remove interference points in radar pulse signals,enhancing sorting precision.Then,it incorporates particle swarm optimization with a silhouette coefficient-based fitness function to adaptively determine optimal clustering parameters.Finally,by integrating a K-Dimensional Tree(K-D Tree),the algorithm significantly reduces the computational complexity of DBSCAN,decreasing sorting time.Experimental results demonstrate that PSK-DBSCAN achieves a sorting accuracy of 98.9% and exhibits robust performance with complex radar signal patterns. |
| Key words: radar signal sorting data field particle swarm optimization K-D tree density clustering |
