基于改进三线性分解的单基地十字阵MIMO雷达二维角度估计

研究单基地十字阵多输入多输出(multiple-input multiple-output,MIMO)雷达中目标二维角度参数估计的问题。已有的算法往往忽略了信源矩阵中的类Vandermonde结构,而这种特殊的结构可以提升参数估计精度。基于均匀线形阵列(uniform linear array,ULA)的中心对称特性和目标参数矩阵中的类Vandermonde结构,提出一种基于改进的三线性分解的二维角度估计算法。首先利用酉变换的方法构造阵列增广输出矩阵,再将二维角度估计与三线性模型相联系。由于增广输出使得阵列的虚拟孔径增大,因而本文所提算法的参数估计精度要优于传统三线性估计算法。此外,本文提及的改进算法不需进行谱峰搜索及奇异值分解,并且能对估计的二维目标角度自动配对,最后的仿真结果验证了本文算法的有效性。

Two-dimensional angle estimation for monostic cross MIMO radar using improved trilinear decomposition algorithm

This research stresses the problem of the two-dimensional angle estimation in monostic cross multiple-input multiple-output (MIMO) radar. Parameters estimation accuracy could be improved with the Vandermonde-like structure in the source matrix, which is always ignored by existing estimation algorithms. With the centrosymmetric of the uniform linear array （ULA） and the Vandermonde-like structure of the data model, an improved trilinear decomposition algorithm is proposed for the two-dimensional angle estimation. The unitary transform is used to construct an expand data matrix, and the two-dimensional angle estimation is then linked to the trilinear model. Due to the expand output, the virtual aperture of the array is increased, hence the proposed trilinear algorithm performs better than the traditional trilinear algorithm. In addition, the proposed algorithm requires neither peak searching nor eigenvalue decomposition. Furthermore, the proposed algorithm could achieve automatic pairing of the two-dimensional angle. Simulation results verify the effectiveness of the proposed algorithm.