基于频繁项集挖掘的雷达PRI模式提取方法

Radar PRI Pattern Extraetion Method Based on Frequent Itemset Mining

针对利用雷达侦察信号挖掘雷达脉冲重复间隔(PRI)模式问题,提出了一种基于频繁项集挖掘的雷达PRI模式提取方法。首先,对PRI值进行离散化聚簇分析,将杂乱的PRI数据变成一组PRI特征值,并按照时间、方位及载频进行聚类分析,从而将同时侦测到的数据归为一类;然后,按类分别提取PRI特征值,得到以PRI特征值为元素的频繁项集;最后,对PRI特征进行关联提取,挖掘出雷达PRI模式。在Spark集群中仿真分析表明,该方法能够从海量数据中并行、高效和准确提取雷达辐射源的PRI特征参数模式。

Aimed at the problem about mining the radar pulse repetition interval （PRI） pattern from the radar reconnaissance signals, a radar PRI pattern extraction method based on frequent itemset mining is proposed. Firstly, the PRI values are analyzed in discretization and clustering. The unorderly PRI data are turned into a group of PRI eigenvalues. The clustering analysis is car tied out according to the time, the azimuth and the frequency. Thus, the data measured in the same time are classified into the same type. Then, the PRI values are extracted by the type re spectively, and the frequent itemset is obtained with the PRI values as the elements. Finally, the association extracting is carried out to the PRI {eature, and the radar PRI pattern is mined. The simulation analysis in the Spark cluster shows that the method can parallel, efficiently and accu rately extract the PRI characteristic parameter pattern of the radar emitters from the massive data.