基于稀疏与低秩矩阵分解理论的广域监视GMTI研究

The urgent demand for Wide-Area Surveillance(WAS) Ground Moving Target Indication (GMTI) system in martial reconnaissance should be fulfilled. The common GMTI methods have disadvantages which result in worse processing performance. It is the first time around the world that the sparse and low-rank matrix separation theory is applied systemically in the multi-channel WAS GMTI radar system in this item. After studying deeply the matrix separation theory and practical radar systems, the mapping relationship between the radar echo model and the matrix separation model is set up. Then the low-rank matrix of the ground clutter, the sparse matrix of the moving targets and the noise matrix are constructed based on mathematical methods. The precise signal transmitting and receiving model is established according to the radar system theory, and the channel imbalance problem is solved by the signal processing theory which is helpful to get the more effective low-rank matrix of the ground clutter and the sparse matrix of the moving targets. Furthermore, the low-rank character of the ground clutter and the sparse character of the moving targets would be thoroughly studied. Associated with the feature of the wide-area surveillance radar system, the mathematical model of the weighting factor is established using the method of theoretical analysis and simulation verification, thus, the uncertainty problem when choosing the weighting factor is solved. Moreover, four matrix separation schemes which are fit for the multi-channel WAS GMTI are proposed based on the existing matrix separation theory. Finally, the matrix separation techniques for the multi-channel WAS GMTI system are developed by constructing the simulation experimental system and are validated by real data processing results.

我国航空航天军事侦察对广域监视(WAS)运动目标指示(GMTI)有迫切需求。传统GMTI方法由于运算量大、对信杂比依赖性较强、易受实际环境杂波非均匀性影响等原因，其处理性能和理想要求具有一定的差距。本项目是国内外首次系统地开展基于稀疏与低秩矩阵分解理论的广域监视运动目标指示研究。依据矩阵分解理论，结合实用化的雷达系统，获得雷达回波模型与矩阵分解模型的映射关系，构建出低秩的杂波矩阵、稀疏的运动目标矩阵和噪声矩阵；通过研究雷达系统信号处理理论，建立精确的信号收发通道模型，解决通道不均衡问题；深入研究杂波低秩机理和运动目标稀疏机理，采用理论分析和仿真验证的方法，建立选取松弛因子的数学模型，以解决现有松弛因子选取的不确定问题；在现有稀疏与低秩矩阵分解技术的基础上，提出四种适合多通道广域监视GMTI的矩阵分解方法。通过构建仿真实验系统，完成原理性验证；通过实际机载飞行数据，进一步完成实用性验证。

本项目是国内外首次系统地开展基于稀疏与低秩矩阵分解理论的广域监视(WAS) 运动目标指示(GMTI)研究。本项目深入研究了稀疏与低秩矩阵分解理论，对多通道广域监视回波的低秩和稀疏机理进行了分析；提出了四种矩阵分解方法，通过仿真数据分析并验证每一种矩阵分解技术的可行性并形成方案；提出了基于修正稀疏特性的矩阵分解方法，并将该方法应用到实际数据中；对关键技术问题进行了深入分析，包括通道均衡、松弛因子、盲速和动目标参数估计等问题，提出了有效的解决办法，通过仿真和实际数据对这些关键技术进行了验证；在现有稀疏与低秩矩阵分解算法的基础上，研究了该算法在常规SAR/GMTI系统中的应用情况，并设计了初步的检测方案。本项目的完成对GMTI领域的发展具有一定的促进意义。