基于分形理论的电台个体识别技术

发布时间：2018-01-20 13:22

本文关键词： 个体特征希尔伯特-黄变换分形维数分类器　出处：《西安电子科技大学》2014年硕士论文　论文类型：学位论文

【摘要】：通信电台个体识别技术作为通信对抗领域的一个重要研究课题,受到国内外研究者的广泛关注。由于不同电台内部器件的细微差异,导致其以幅度、频率和相位的表现形式附加在发射信号上。电台个体识别技术就是通过从截获信号中提取能表征电台身份的特征信息,判断信号来自哪一个电台,从而实现对电台的分析,跟踪,定位以及精确打击。本文在深入分析电台个体特征的产生机理之后,针对电台在稳态情况下,采用基于希尔伯特-黄变换(HHT)边际谱的分形特征方法提取信号在频率方面表现出的个体特征。该方法首先采用基于经验模态分解(EMD)的希尔伯特-黄变换(HHT)算法提取信号的边际谱,然后采用分形维数提取边际谱中附带的特征信息。该方法能够较好的解决现有的谱对称算法中存在的特征点分布聚集度较差以及精度不高,时效性较差等问题。实验结果表明采用基于希尔伯特-黄变换(HHT)边际谱的分形特征方法能够获得良好的特征点聚集性。当边际谱的分辨率降低时,仍然能够达到较好地识别效果,并且能够在一定程度上提高算法的时效性。最后,本文采用了基于OAO算法的多类别支撑向量机分类器对提取的特征向量进行分类,实验结果证明该分类器能够达到较好的效果。
[Abstract]:As an important research topic in the field of communication countermeasure, individual identification technology of communication radio station has been widely concerned by researchers at home and abroad. Due to the subtle differences in internal devices of different radio stations, it is caused by the amplitude. The expression of frequency and phase is attached to the transmitted signal. The individual identification technology of the radio station is to extract the characteristic information which can represent the identity of the station from the intercepted signal to determine which station the signal comes from. In order to achieve the radio analysis, tracking, positioning and accurate attack. In this paper, after in-depth analysis of the radio station individual characteristics of the mechanism, the radio station in the steady state. A fractal feature method based on Hilbert-Huang transform (HHT) marginal spectrum is used to extract the individual characteristics of signals in frequency. The method is based on empirical mode decomposition (EMD). The Hilbert-Huang transform (HHT) algorithm extracts the marginal spectrum of the signal. Then the fractal dimension is used to extract the feature information attached to the marginal spectrum. This method can solve the problem of poor clustering degree and low accuracy of the existing spectral symmetry algorithm. The experimental results show that the fractal feature method based on Hilbert-Huang transform (HHT) marginal spectrum can obtain good feature point aggregation when the resolution of the marginal spectrum is reduced. Still can achieve a better recognition effect, and to a certain extent can improve the timeliness of the algorithm. Finally. In this paper, a multi-class support vector machine classifier based on OAO algorithm is used to classify the extracted feature vectors. The experimental results show that the proposed classifier can achieve good results.
【学位授予单位】：西安电子科技大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN975

【相似文献】