基于改进K-均值算法的未知雷达信号分选

针对K-均值算法需要事先确定聚类的数目,无法适用于未知雷达信号分选的问题,通过引入脉冲间欧几里德距离和距离阈值TMS2812,完成聚类数目和聚类中心的自动选取,给出一个K-均值的改进算法,改进后的算法既收敛速度快,易于工程化实现,又可自动确定聚类数目和聚类中心。仿真实验表明,该改进算法提高了K-均值算法的适用范围,能够有效适应于未知雷达信号的分选。

Unknown Radar Signals Deinterleaving Based on Improved K-means Algorithm

Since K-means algorithm can not be used for the deinterleaving of the unknown radar signals and needs to make sure the clustering number beforehand, an improved K-means algorithm is proposed based on the introduction of Euclidean distance and the distance threshold value, and the achievement of the automatic selection for the clustering number and clustering center. The algorithm can keep the original advantage and doesn＇t need the clustering number. The simulation shows that this algorithm enlarges the application range of K-means algorithm and can successfully deinterleave unknown radar signals.