基于瞬时频率二次特征提取的辐射源信号分类

提出了基于瞬时频率二次特征提取的雷达辐射源信号分类方法．首先利用改进的瞬时自相关算法提取信号的瞬时频率特征．在此基础上，对所获得的瞬时频率进行级联归一化处理，提取分类特征向量．最后，采用层次决策方法实现自动分类．仿真结果表明，该方法提取的特征向量具有较好的类间分离性，整体信号分类方案在信噪比不低于6dB时，可获得90％以上的分类准确率．

Automatic Classification of Radar Emitter Signals Based on Cascade Feature Extractions

A scheme of automatic classification of radar emitter signals was presented, which is based on cascade feature extractions. First, an improved instantaneous autocorrelation algorithm is performed to extract instantaneous frequencies of typical radar emitter signals. Then, the extracted instantaneous frequencies are normalized twice to obtain the classification characteristic vector. Finally, the hierarchical decision approach is used to classify radar signals automatically. The results of simulation show that the classification characteristics vector has good separation property between clusters, and the proposed approach achieves a correct rate of above 90%, even when the slgnal-to-noise ratio is as low as 6 dB.