基于Nyström的一种新颖的DOA估计算法

A Novel DOA Estimation Algorithm Based on Nyström

针对阵列信号处理领域中的超分辨子空间类算法需计算阵列输出的协方差、协方差矩阵的特征分解及进行谱峰搜索得到波达方向(Direction of Arrival,DOA)估计,计算量较大则实际应用可能受限,提出了一种低计算复杂度的新颖的无需谱峰搜索的DOA估计算法。通过Nyström方法得到逼近的信号子空间,避免了直接对所有阵列输出计算协方差及对其特征分解从而降低了运算量。通过逼近的信号子空间构建低阶的关于DOA的特征多项式方程,对此低阶多项式方程求根得到DOA估计进一步降低了运算量,且不同于现有的DOA估计求根算法。理论分析和仿真结果表明,所提算法有着良好的估计精度及较低的计算复杂度。

The super-resolution subspace algorithms in the field of array signal processing need to calculate the covariance of array output,the eigen-decomposition of the covariance matrix of the array output and perform spectral peak search to obtain direction of arrival(DOA)estimation,which may be limited in practical application due to the high computational complexity.A novel DOA estimation algorithm with low computational complexity and no need for spectral peak search is proposed.The approximate signal subspace is obtained by Nyström method,which avoids directly calculating the covariance and its eigen-decomposition for all the array outputs,thus reducing the computational complexity.Then,a DOA-dependent low-order characteristic polynomial equation is constructed through the approximate signal subspace,and the DOA estimation is obtained by rooting this low-order polynomial equation,which further reduces the computational complexity and is different from the existing DOA estimation algorithms based on polynomial root.Theoretical analysis and simulation results show that the proposed algorithm has good estimation accuracy and low computational complexity.