基于矩阵秩显QR分解的时延估计

近年来，ＭＵＳＩＣ算法在时延估计研究中异常活跃，它获得和利用接收信号的协方差矩阵的一组与信号时延矢量正交的特征矢量。特征分解需要很大的计算量。本文基于矩阵秩显ＱＲ分解提出一种新的时延估计方法。计算机模拟结果表明此方法与ＭＵＳＩＣ算法同样有效，而且计算量大为减小。

Estimation of Time Delay Based on Matrix Rank Recealing QR Factorization

In recent years, MUSIC algorithm is frequently adopted in the estimation of time delay.It obtains and uses a set of vectors orthognal to signal time delay vectors by performing an eigendecomposition of covarince matrix of veceiving signal. Eigendecom position, however, requires a considerable amount of computations. Based on a permuted matrix rank revealing QR facterization,a new method is presented in this paper for the estimation of time delay. Computer simulation shows that the proposed method can perform almost as well as MUSIC algorithm in terms of signal resolition,and the cost of this method is much lower than eigendecomposition.