一种快速的支持向量聚类雷达信号分选方法

摘要

Since the existing support vector clustering(SVC)algorithm has high complexity,which can not meet both the real-time and accuracy requirements in the dense,complicated and changefully modern radar countermeasure signal environments, a new radar signal sorting method based on the cone reflection SVC algorithm is proposed. The clustering relust was verified by clustering factor regulation of entropy feature. The simulation results show that the method can improve the radar signal sorting velocity,while ensuring the correctness of signal sorting. It is of a certain practical value.%针对现有支持向量聚类算法复杂度较高，在密集、复杂、多变的现代雷达对抗信号环境中，无法同时满足雷达信号分选的准确性和实时性，提出了锥面映射支持向量聚类的雷达信号分选方法，并利用熵表征的聚类因子调整验证聚类效果。仿真结果表明，该方法在保证信号分选正确率的同时，可有效提高雷达信号分选速度，具有一定的实用价值。