辐射源信号指纹识别技术

发布时间：2018-02-12 06:31

  本文关键词： 辐射源信号指纹识别 相位噪声 载频估计 杂散特征 高阶谱特征　出处：《电子科技大学》2014年硕士论文　论文类型：学位论文

【摘要】：辐射源指纹识别技术是信号处理学科中一门新兴的研究方向。因此,本文通过对通信信号辐射源的个体细微特征进行分析与提取,并利用这些特征进行分类识别技术进行了研究。本文的主要工作包括以下3个方面:1.研究了基于码速率和载频特征提取技术,改进了相位差分载频估计算法,利用前后向相位差分方法,对噪声相位增量做两次平滑处理,降低了相位噪声对载频估计的影响,提高了估计精度。并根据Haar小波变换法与短时傅立叶变换法的时频能量对码速率估计方法进行了研究。理论分析与仿真实验表明,本文的码速率和载频估计精度能够满足对截获的电台信号载频特征提取的要求。2.研究了辐射源包络信息杂散特征的分析与提取技术。给出了Hilbert变换法、Teager-Kaiser法以及基于信号正交分量重构的包络提取算法;研究了信号包络的R与J特征、描述分形集的复杂性和不规则性的盒维数特征、描述分形集在区域空间上的分布疏密的信息维数特征以及描述信号包络变化规律的Lempel-Ziv复杂度特征等包络调制特征提取与分析方法。然后利用截获信号的Hilbert边缘谱谱对称特性和Hilbert Huang变换时频分布灰度图像提取信号的指纹特征。实验表明不同信号频谱对称性特征向量间的差别较大,聚类效果较好,适用于识别分类不同的辐射源信号。3.研究了基于高阶谱的信号指纹分析与提取。通过对四种局部双谱,以及矩形积分双谱等基于高阶谱的信号指纹特征提取方法,并利用Fisher鉴别函数仿真验证了这些双谱分析方法在信号特征提取方面的性能。实验表明,矩形积分双谱相较局部双谱分析方法,在提取信号个体特征方面的性能较好。
[Abstract]:The fingerprint identification of emitter is a new research direction in the field of signal processing. Therefore, this paper analyzes and extracts the individual fine features of the emitter of communication signal. The main work of this paper includes the following three aspects: 1.The technology based on code rate and carrier frequency feature extraction is studied, and the phase difference carrier frequency estimation algorithm is improved. The phase increment of noise is smoothed twice by the method of forward and backward phase difference, and the influence of phase noise on carrier frequency estimation is reduced. The estimation accuracy is improved. The method of code rate estimation based on Haar wavelet transform and short time Fourier transform is studied. The theoretical analysis and simulation results show that, The code rate and carrier frequency estimation accuracy of this paper can meet the requirements of carrier frequency feature extraction of intercepted radio signals. 2. The analysis and extraction techniques of stray features of emitter envelope information are studied. The Hilbert transform method and the base method are given. The envelope extraction algorithm based on orthogonal component reconstruction; The R and J characteristics of the signal envelope are studied, and the box dimension features of the complexity and irregularity of the fractal set are described. The method of extracting and analyzing the information dimension features of fractal set and the Lempel-Ziv complexity feature describing the variation of signal envelope is presented. Then, the Hilbert edge spectrum of intercepted signal is used. Spectral symmetry and fingerprint feature extracted from time-frequency gray-scale image of Hilbert Huang transform show that the difference of spectrum symmetry characteristic vectors between different signals is great. The clustering effect is good, which is suitable for the recognition of emitter signals with different classification. The fingerprint analysis and extraction of signals based on high order spectrum are studied. As well as rectangle integral bispectrum and other signal fingerprint feature extraction methods based on high order spectrum. The simulation of Fisher discriminant function is used to verify the performance of these bispectral analysis methods in signal feature extraction. The experimental results show that, Compared with the local bispectrum analysis method, the rectangular integral bispectrum has better performance in extracting individual features of signals.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN911.7

【参考文献】

中国博士学位论文全文数据库 前1条

1 张国柱;雷达辐射源识别技术研究[D];国防科学技术大学;2005年

，

本文编号：1504998