基于双谱特征的个体辐射源识别

Specific emitter identification based on bispectrum feature

针对个体辐射源识别困难的问题,提出了一种基于双谱分析的个体辐射源识别新方法。首先用双谱分析法提取信号的双谱对角切片,利用主成分分析法(PCA)从大量训练样本特征中挑选低维、低复杂度的特征矢量,并融合对分类具有显著贡献的辐射源属性参数作为识别特征矢量。最后采用势函数分类法实现雷达辐射源识别。仿真结果表明基于双谱的识别法能较好地解决同型号、同标称的雷达辐射源个体识别问题,其识别率高达93%。

With respect to specific emitter identification, a new identification approach for radar emitter based on bispectrum analysis is proposed. Diagonal slice features of bispectrum which adapts to process radar non-linearity signals are extracted, and principal analysis is utilized to extract low-dimensional feature vectors with low complexity from a large of training sample features. Moreover, emitter attribute parameters significant to identification are merged into the identification feature vector. At last, radar emitter identification is completed with potential function classification method. Simulations results show that the method can solve the problem of identifying same model radar emitter with a 93 % identification rate.