一种基于小波和分形理论的电台个体识别方法  

Individual Transmitter Identification Based on Wavelet and Fractal Theory

作　　者：赵国庆[1,2] 彭华[1] 王彬[1] 丁金忠[3] 

机构地区：[1]信息工程大学信息工程学院 [2]72959部队 [3]95851部队

出　　处：《信息工程大学学报》2012年第1期76-81,共6页Journal of Information Engineering University

基　　金：国家自然科学基金资助项目(61072046)

摘　　要：从调制信号中提取发射电台的频率稳定度特征进而实现个体识别是一项极具挑战性的课题。针对MPSK信号提出了一种基于小波和分形理论的特征提取方法。该方法首先利用连续小波变换把MPSK信号幅度中的振荡器频率稳定度信息转移到小波系数中,然后从统计角度提取小波系数的盒维特征,最后用K-最近邻分类器进行个体区分。仿真实验表明,在10dB～20dB内,对于载波频率稳定度差异为ppm级的MPSK信号,该算法平均分类准确率可达95%,为通信电台个体识别技术的实用化提供了进一步理论依据。It＇ s challenging to extract oscillator frequency stability characteristics of a communication transmitter from modulated signals and then realize individual communication transmitter identification. A method based on wavelet transformation and fractal theory is presented for MPSK signals. Information of oscillator frequency stability contained in MPSK signals is transferred to wavelet coefficient using continuous wavelet transformation. Then the box-counting dimension features of the wavelet coefficient are extracted based on statistical data. And finally the communication transimitters are identified with the k-nearest neighbor classifier. Simulation results indicate that this method＇ s accuracy is up to 95% for MPSK signals transmitted by disparate communication transmitters whose frequency stability is of ppm level. This study provides theoretical basis for real-world application of individual communication transmitter identification.

关 键 词：电台个体识别 频率稳定度 连续小波变换 盒维 

分 类 号：TN971.1[电子电信—信号与信息处理]