一种DOA估计的快速子空间算法

A Fast Subspace Algorithm for DOA Estimation

MUSIC算法是一种属于特征结构的子空间超分辨方法,该算法性能优良,但需要估计协方差矩阵并对其进行特征分解,运算量大,很费计算时间。本文对波达方向估计问题进行了研究并提出了一种采取降维处理的快速子空间算法,该算法利用阵列协方差矩阵的一个子矩阵快速得到信号子空间,无需特征分解,且无需估计整个协方差矩阵,只需估计该子矩阵,故快速算法运算复杂度远低于MUSIC算法,而性能损失并不太大。理论分析和计算机仿真结果表明此方法是有效的。

The MUSIC algorithm that belongs to a subspace method of eigenstructure is a superresolution method and it has good performance. However, the estimation of the complete covariance matrix and the subsequent eigendecomposition are computationally intensive and time consuming. The problem of directionofarrival (DOA) estimation is studied and a fast subspace algorithm using reduced dimension processing is proposed. This algorithm can fast obtain the signal subspace from a submatrix of the array covariance matrix without eigendecompsition and it only needs estimating the submatrix instead of estimating the whole covariance matrix. The algorithm has a low computational complexity in comparison with the MUSIC algorithm without sacrificing too much performance. Theoretical analysis and computer simulation results show that the method is effective.