基于杂草改进的模糊聚类雷达信号分选

针对雷达辐射源信号参数严重混叠、聚类数目未知等问题,提出一种基于入侵性杂草优化模糊聚类的智能算法,该算法无需事先设定聚类数目,而是在整个数据集的属性空间内并行搜寻最佳的聚类数目和聚类中心,具有结构简单、鲁棒性好的特点。将此方法应用到雷达信号的分选当中,并与传统的K均值算法及AP聚类算法进行对比,实验结果验证了该算法的有效性。

Radar signal sorting based on the improvement of fuzzy clustering by weeds algorithm

In the condition of unknown clustering number, this paper puts forward a kind of intelligent algorithm based on invasive weeds algorithm to optimize fuzzy clustering for solving the problem of seriously aliased parameters of radar emitter signals. This algorithm searches parallel optimal cluster number and cluster center in the space of properties of entire data set, instead of setting the cluster number beforehand, with characteristics of simple structure and good ro-bustness. Applying the method to the radar signal deinterleaving and comparing it with the traditional k-means algo-rithm and AP clustering algorithm, the experimental results verify effectiveness of the algorithm.