基于欠定盲分离的跳频信号分选和识别技术研究

项目名称： 基于欠定盲分离的跳频信号分选和识别技术研究

项目编号： No.61201134

项目类型： 青年科学基金项目

立项/批准年度： 2013

项目学科： 电子学与信息系统

项目作者： 付卫红

作者单位： 西安电子科技大学

项目金额： 26万元

中文摘要： 盲源分离技术由于其盲的特性，在通信对抗中具有广泛的应用前景。针对欠定混合条件下多个跳频信号的分选和识别技术，利用跳频信号频域的稀疏性，研究基于欠定盲分离的跳频信号分离技术，包括时变欠定混合矩阵的估计以及跳频源信号的恢复。根据充分稀疏信号的直线聚类或非充分稀疏信号的平面聚类特性，利用自适应聚类算法以及势函数的思想，提出一种新的适用于时变信道的欠定混合矩阵实时估计算法；然后针对非充分稀疏源信号的恢复问题，提出一种稳健的欠定混合跳频源信号恢复算法；最终给出一套完整的欠定混合跳频信号分离算法，该算法计算复杂度低、精确度高、抗噪能力强。在此基础上，对分离出来的单一跳频信号的参数识别技术展开研究，利用压缩感知理论，提出基于压缩感知的跳频信号参数（跳频频率、跳频速率、跳变时刻）识别算法。本项目的研究为跳频信号的侦查开辟一条新的技术途径。

中文关键词： 跳频信号分选；欠定盲分离；混合矩阵估计；压缩感知；稀疏信号重构

英文摘要： Blind source separation technique due to its blind nature, have a wide range of applications in communication confrontation. Aiming to the multiple frequency hopping (FH) signal separation and identification technology for underdetermined mixed conditions, by using the sparsity of FH signal in Frequency domain, we research the separation technology of the FH signal, including estimation of the Time-varying underdetermined mixing matrix and the restoration of FH source signals. According to the line clustering properties of the full sparse source and plane clustering properties of the non-sufficient sparse source, by using the idea of adaptive clustering algorithm and potential function, a novel real-time underdetermined mixing matrix estimation algorithm adapt to time-varying channel will be presented. Aiming to the problem of the restoration of the non-sufficient sparse source signal, a robust restoration algorithm for underdetermined mixed FH signal will be proposed. Finally we give a complete separation algorithm of underdetermined mixed FH signal with low computational complexity, high accuracy, and strong anti-noise. On this basis, parameters identification technology of a single separated FH signal is researched by using compressed sensing theory, and a FH signal parameters (Hoping frequencies, hopping rat

英文关键词： frequency hopping signal sorting；Underdetermined blind source separation；Mixing matrix estimation；compressed sensing；sparse signal reconstruction