基于平行因子压缩感知理论的阵列信号处理算法研究

This project is to extend compressive sensing theory to tensor signal processing problem, and proposes the parallel factor (PARAFAC) compressive sensing framework, which compresses PARAFAC model to low dimensional PARAFAC model and then recoveries signal using the signal sparisty. It can greatly reduce complexity and storage of the PARAFAC decomposition. This project applies the theory of PARAFAC compressive sensing to array signal processing, and the concrete research of this project is mainly focused on: ① PARAFAC compressive sensing framework design. ② The array signal detection with PARAFAC compressive sensing. ③ The array parameter estimation with PARAFAC compressive sensing. ④ The array signal processing with PARAFAC completion. ⑤ The array signal processing with PARAFAC compressive sensing for coherent source signals and multipath channel. Applying the theory of PARAFAC compressive sensing to array signal processing, which aims at solving the key technique of array signal processing, as well as making breakthroughs on the limits of traditional methods for array signal processing, will provide a new method for both the signal detection and high resolution parameter esimation for array signal processing. The theory of PARAFAC compressive sensing and its application to array signal processing will have significant values on theoretical research, and thus have a widely bright future for prospective applications. The project research is undoubtedly of great theoretical and practical significance for our future development of a new generation of information acquisition and transmission system.

本课题将压缩感知理论推广到高阶张量信号处理，提出了平行因子压缩感知理论的框架，将平行因子模型压缩到低维平行因子模型，再利用信号稀疏性来恢复信号，这样可大大减小平行因子分解复杂度和存储量。开展基于平行因子压缩感知理论的阵列信号处理算法研究，具体研究内容如下：①平行因子压缩感知理论的框架设计，②基于平行因子压缩感知的阵列信号检测研究，③基于平行因子压缩感知的阵列参数估计研究，④基于平行因子模型填充的阵列信号处理研究，⑤相干信源和多径信道下基于平行因子压缩感知的阵列信号处理研究。本项目旨在攻克平行因子压缩感知理论中的关键技术，并突破传统阵列信号处理方法中的局限，为阵列信号检测和高分辨参数估计提供一种新方法。开展本项目的研究对于自主发展我国未来新一代信息获取与传输系统无疑具有十分重要的理论与实际意义。

本课题将压缩感知理论推广到高阶张量信号处理，提出了平行因子压缩感知理论的框架，将平行因子模型压缩到低维平行因子模型，再利用信号稀疏性来恢复信号，这样可大大减小平行因子分解复杂度和存储量。开展基于平行因子压缩感知理论的阵列信号处理算法研究，具体研究内容如下：①平行因子压缩感知理论的框架设计，②基于平行因子压缩感知的阵列信号检测研究，③基于平行因子压缩感知的阵列参数估计研究，④基于平行因子模型填充的阵列信号处理研究，⑤相干信源和多径信道下基于平行因子压缩感知的阵列信号处理研究。在该项目支持下，课题组申请发明专利4项，培养研究生和博士生15 名，出版专著2部，发表论文40多篇，其中SCI检索论文20多篇。本项目攻克平行因子压缩感知理论中的关键技术，并突破传统阵列信号处理方法中的局限，为阵列信号检测和高分辨参数估计提供一种新方法。开展本项目的研究对于自主发展我国未来新一代信息获取与传输系统无疑具有十分重要的理论与实际意义。