一种修正的矢量奇异值分解DOA估计算法

在均匀线性阵列模型下，特征矢量奇异值分解算法能够对相干信号进行DOA估计， 但相干和不相关信号同时存在时，算法的估计会出现错误。针对这一问题，提出了一种修正 算法（MESVD），该算法选取经过加权处理的特征向量来构造矩阵，并利用该矩阵进行子空 间估计。理论分析和数值仿真证明：修正后的算法能够正确估计相干、相关和不相 关信号， 估计性能与空间平滑算法（FBSS）相当。

A modified ESVD algorithm for DOA estimation

The Extended Singular Value Decomposition(ESVD) algorithm can deal with the coh erent signals exactly in Uniform Linear Array(ULA), but when the coherent signa ls and the uncorrelated signals inject at the same time, the ESVD algorithm usua lly gives incorrect result. According to this problem, a modified algorithm(MESV D) is proposed, which uses a weighted eigenvector to construct a matrix for sub spaces estimation.Analysis and simulations show that the MESVD algorithm can giv e right e stimation without considering the relativity of signals and its estimation perfo rmance is corresponding to that of FBSS algorithm.