雷达辐射源无意调制特征提取算法研究

发布时间：2018-01-14 18:09

本文关键词：雷达辐射源无意调制特征提取算法研究　出处：《电子科技大学》2014年硕士论文　论文类型：学位论文

【摘要】：雷达辐射源的个体识别是为了适应雷达新技术和日益复杂的电磁环境需发展起来的,是电子对抗领域的重点发展方向。无意调制特征的提取是个体识别系统中的关键技术之一。不同于信号的基本特征,无意调制特征来源于发射机内部器件尤其是功率放大器的非线性特性,它具有唯一性、普遍性、独立性、可测性等特点。本文围绕雷达辐射源的无意调制特征提取展开研究,从发射机无意调制特性的机理入手,研究了三种无意调制特征的提取算法。论文的主要工作包括:1.研究了发射机的非线性调制特性,分析了无意调制特征产生的机理。针对功率放大器是无意调制的主要来源,详细分析了其对辐射源信号时域和频域的影响,并建立了数学模型,作为后续章节研究的基础。2.针对个体特征抗干扰能力及区分度差等问题,研究了基于模糊函数和小波变换的无意调制特征提取算法。利用模糊函数将辐射源信号变换到时延-频偏二维平面,提取信号的斜切片特征并通过小波变换获得这种特征的小波系数,采用改进的Fisher线性判别进行特征降维以得到有区分度的个体特征。仿真和实测数据结果验证了所研究算法的有效性。3.针对高阶统计量具有抑制噪声、保留信号相位等特点,研究了基于高阶统计量特征的无意调制特征提取算法。通过双谱变换提取最大谱峰的峰值和相位、幅度和相位垂直切片积分谱的均值特征并提取了信号的偏度和峭度特征,仿真和实测数据结果验证了所研究算法的有效性。4.针对时间序列是高维动力系统在一维空间映射的观点,研究了基于相空间和GMM(Gaussian mixture model:混合高斯模型)的无意调制特征提取算法。在重构辐射源信号相空间的基础上,提取相空间的质心特征,以此为对象建立GMM模型并计算对应的似然值进行分类识别。仿真和实测数据结果验证了所研究算法的有效性。同时,对比分析了本文所提三种算法在不同信噪比下的识别性能。
[Abstract]:Individual identification of radar radiation source is in order to adapt to the new technology of radar and increasingly complex electromagnetic environment should be developed, is a key development direction in the field of electronic warfare. The extraction of unintentional modulation feature is one of the key technologies of individual identification system. The signal is different from the basic characteristics, the unintentional modulation feature comes from inside the transmitter device especially the nonlinear characteristic of power amplifier, it has the universality, uniqueness, independence, scalability and other characteristics. This paper focuses on the radar radiation source of unintentional modulation feature extraction is studied, starting from the transmitter modulation characteristics of unintentional mechanism, three kinds of unintentional modulation feature extraction algorithm. The main work includes: 1. the study of the nonlinear modulation characteristics of the transmitter, analyzes the mechanism of unintentional modulation feature generation. For power amplifier is the main source of unintentional modulation, detailed analysis The influence on the radiation signal in time domain and frequency domain, and establishes a mathematical model, the.2. problem as the basis of the following chapters according to the individual characteristics of the indexing anti-interference ability and poor area, studies the extraction algorithm of unintentional modulation feature based on wavelet transform and fuzzy function. Using the fuzzy function of the radiation source signal to delay - frequency plane, inclined slice wavelet coefficients of signal feature extraction and the characteristics of the wavelet transform, using Fisher linear discriminant feature dimension reduction to obtain the individual characteristics of discrimination. Simulation and experimental results verify the validity of the.3. data of the proposed algorithm for high-order statistics can suppress noise retention characteristics of signal phase, based on the higher-order statistics characteristic of unintentional modulation feature extraction algorithm. By bispectrum transform to extract the maximum spectral peak and peak phase, The mean amplitude and phase spectrum characteristics of integral vertical slice and pick up the signal of skewness and kurtosis characteristics, simulation and experimental data to test the effectiveness of.4. algorithm for time series is a point in one dimensional space mapping high dimensional dynamical system, the research of phase space based on GMM and Gaussian (mixture model: mixed Gauss model) unintentional modulation feature extraction algorithm based on phase space reconstruction. The radiation source signal and centroid feature extraction phase space as the object, established the GMM model and the calculation of the corresponding likelihood value classification. Simulation and experimental data results verify the effectiveness of the proposed algorithm in this paper. At the same time, analyzed in this paper three proposed algorithms under different SNR recognition performance.

【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】：TN974

【相似文献】