复杂电磁环境下单天线接收时频混叠信号分离方法研究

In nowadays information era, the types of communication or radar electronic equipment for both military and civilian applications are increasing significantly, which leads to various communication and radar signals overcrowded in time domain, overlapped in frequency domain and intertwined in space domain shaping a more complicated electromagnetic environment. As a result, the interception probability of time-frequency overlapped signals has been improved for electronic reconnaissance equipment. In order to accurately capture the interested signal or discover the interference signal, separating these time-frequency overlapped signals and extracting the information implied in the useful signal have become a research task with significant for electromagnetic surveillance domain. Meanwhile, the electronic reconnaissance equipment with single antenna not only reduced the cost but can reduce the equipment volume and improve its flexibility. Taking this background in to consideration, this project focuses on time-frequency overlapped signal separation theory and method research for single-antenna electronic reconnaissance in complex electromagnetic environment. Using the time structural information of the original time-frequency overlapping signals’ oversampling result, it will focus on modeling, identification and reconstruction technique for time-frequency overlapped signal from single antenna. So as to break through the limitation of existing multi-channel signal separation technique for time-frequency overlapped signals, and to provide effective separation and information capture theory and methods for time-frequency overlapped signals interpreted by reconnaissance equipment in actual complex electromagnetic environment.

在当今信息化时代，通信、雷达等各种军用和民用电子设备的广泛应用，形成了时域高度密集、频域严重重叠、空域相互交织的复杂多变电磁环境，电子侦察机侦收时频混叠信号的概率显著提高。为了准确获取感兴趣的信号或发现重大干扰、威胁信号，对侦察机侦收的时频混叠信号进行分离，提取有用信号蕴含信息，成为电子侦察领域具有重要研究意义的项目。同时，单天线侦察机不仅可以降低设备造价，而且能够减小设备体积增强灵活性。在此背景下，本课题针对复杂电磁环境下单天线接收时频混叠信号的分离理论与方法展开研究。利用时频混叠信号在过采样后的时间结构信息，重点攻研单天线接收时频混叠信号的建模、分离和重构技术，突破现有多通道技术对时频混叠信号分离的限制，为侦察装备在实际的复杂电磁环境下对时频混叠信号的有效分离和信息获取提供理论基础和方法。

本项目针对当前时域高度密集、频域严重重叠、空域相互交织的复杂电磁监测环境下信号侦察的需求，围绕单天线接收时频混叠信号的分离问题，开展了信号建模、分离和重构算法研究。首先，在特性分析的基础上，研究通信信号混合、雷达信号混合以及通信信号与雷达信号混合的信号建模技术，包括基于基展开和数据数据拟合的混合信号建模、单通道虚拟多通道的建模，完成了高拟合度单通道多维度重叠信号的时域建模。第二，在单天线接收时频混叠信号特征分析、联合参数估计的基础上，研究单天线接收时频混叠信号分离与重构算法，包括基于单通道虚拟多通道技术的信号分离与重构算法、基于信号特征差异性的信号分离与重构算法、基于源信号特性构建目标函数采用机器学习优化方法完成信号分离与重构的算法。第三研究了分离信号的盲解调、调制类型识别与参数估计方法，用于信号分离精度与性能的评估。此外，项目组基于Matlab语言设计项目配套的仿真软件平台。仿真实验表明，在（Signal-to-Noise Ratio,SNR），源信号频域重叠度（Frequency Overlapped Level）甚至100%时，分离通信信号的误码率（Bit Error Rate）BER优于0.01甚至0.001、雷达信号的信干比（Signal-to-Interference Ratio），性能指标(performance index)。攻克复杂电磁监测环境下单天线接收时频混叠信号的高精度分离技术，从而准确获取感兴趣的信号或发现重大干扰、威胁信号，并提取信号蕴含有用信息，是电子侦察领域重要研究课题。同时，部分信号分离算法可以移植应用于混响环境下多说话人语音信号分离，助力于人工智能领域人机交互技术的发展。