摘要
提出一种以盒维数、信息维数、相像系数为分类特征识别雷达信号调制方式的方法。这些特征包含了信号的幅度、频率、相位、整体走势(或者轮廓)、波形复杂度和不规则度的细节信息,集中体现了不同调制方式的差异。同时,利用特征自身的类内距离小、类间距离大的特点先聚类分离部分调制方式,最后通过基于粗集的支持向量机分类器进一步分类识别。通过4种典型雷达辐射源信号的特征提取与分类识别的仿真试验,表明基于本方法的调制信号识别正确率高,具有一定的工程应用价值。
A method is proposed for modulation recognition of radar signal by the classification characters including box dimension,information dimension,and resemblance coefficient.The characters contain the amplitude,frequency,phase and figure of radar signal,and complexity and irregularity of waveform,which can show the distinctness of different modulations include the transformation rules.The method separates different signal by clustering,and make further identification by using RS-SVM as classifier.Simulation was made for feature extraction and classification/identification of four kinds of typical radar radiation signals,the result showed that the method is effective and is suitable for engineering application.
出处
《电光与控制》
北大核心
2012年第11期31-34,38,共5页
Electronics Optics & Control
基金
国家自然科学基金(60901069)
关键词
电子对抗
雷达信号
调制识别
特征提取
相像系数
分形
electronic countermeasure
radar signal
modulation recognition
feature extraction
resemblance coefficient
fractal
