利用信号围线积分双谱分形特征实现电台识别  

Individual Transmitter Identification Based on Fractal Feature of Surrounding-Line Integral Bispectrum

作　　者：何永亮[1] 陈西豪[1] 许华[1] 江汉[1] 刘潇文 

机构地区：[1]空军工程大学信息与导航学院,西安710077

出　　处：《电讯技术》2014年第10期1354-1359,共6页Telecommunication Engineering

基　　金：国家自然科学基金资助项目(61001111)~~

摘　　要：信号的双谱能反映信号的细微特征,可用于电台识别中,但将它直接应用于电台识别需要计算复杂的匹配模板,增加分类器的复杂度,影响识别效率。针对此问题,提出了一种将信号围线积分双谱的分形特征作为电台特征参数的识别方法。首先由信号双谱估计值求出围线积分双谱,然后利用盒维数和信息维数定量描述围线积分双谱波形的复杂度,并将这两种分形维数作为特征向量,最后利用支持向量机(SVM)实现电台分类识别。对两部实际电台所发射的2FSK信号利用所提方法进行分析,结果表明在信噪比为7 dB及以上时,电台正确识别率能达到94.29%以上,验证了所提方法的可行性。Bispectrum which can be used for transmitter identification reflects the fine features of signals, but it will lead to the work of computing the complex matching templates,adding the complexity of classifier and influencing the recognition efficiency when applied to the transmitter identification directly. To solve this problem,a novel approach treating the fractal features of surrounding-line integral bispectrum as the feature parameters of transmitters is proposed. Surrounding-line integral bispectrum is obtained by the bis-pectrum estimator,and the complexity of its wave shape can be quantitatively described by box dimension and information dimension which are selected as the eigenvector. Then the support vector machine（ SVM） is applied for individual transmitter identification. Finally,the 2FSK signals sent by two different transmit-ters are analyzed,and the results show that the recognition accuracy is over 94. 29% when signal-to-noise ratio（SNR） is more than 7dB,which proves that the introduced method is effective.

关 键 词：个体识别 细微特征 双谱估计 分形特征 围线积分 支持向量机 

分 类 号：TN911.72[电子电信—通信与信息系统]