基于动态网格密度聚类的雷达信号分选算法

摘要

Aiming at the problems that the fixed grid density clustering algorithm has dependence of the data input and the clustering accuracy is low in radar signal sorting,a radar signal sorting algorithm based on dynamic grid density clustering is proposed. In which the dynamic grid generation technique is used to greatly reduce the number of null grid,the shifting grid technique is employed to eliminate the dependence of data input,and the modified double-density threshold is adopted to im-prove the clustering precision. The algorithm can cluster the data of any size and shape without any parameters input. The experi-ment results show that the algorithm can eliminate noises effectively and cluster the unknown radar signal rapidly and successfully.%针对固定网格密度聚类算法在雷达信号分选中对数据输入的依赖性和聚类精度低的问题，提出了一种基于动态网格密度聚类的雷达信号分选算法。该算法采用动态网格生成技术大幅度地减少无效网格生成数目，移动网格技术消除对数据输入的依赖性，用改进的双密度阈值提高聚类精度。该算法不需要输入任何参数就能快速对任意形状大小的数据集进行聚类。实验表明，该算法能有效去除噪声，快速准确的对未知雷达信号进行聚类。