一种基于Householder变换的递归 QRD-LS算法

本文提出了一种采用Householder变换实现的递归QRD-LS算法,该算法通过采用Householder变换取代Giv- ens旋转递归实现复矩阵的QR分解来求解LS问题,可以获得比基于Givens旋转的QRD—LS算法更快的处理速度。此外, 算法引入了复数QR分解,解决了算法只能处理实数信号的问题。通过定义新的数据矩阵,算法还可以合并求解数据域正规方程中的系数矩阵和右侧向量,从而提高了计算效率。通过对其在智能天线中的应用进行仿真,验证了算法的性能。

A Recursive QRD - LS Algorithm Based on Householder Transformation

A complex QRD - LS algorithm based on Householder transformation is devised. By substituting conventional Givens rotation with Householder transformation significantly faster speed can be obtained. Moreover, it can solve the LS problem effectively in complex signal scenarios, to which the real QR decomposition approach cannot be applied directly. Using complex QR decomposition and Householder transformation, the algorithm can be recursively implemented. By defining a new data matrix the coefficient matrix and the dexter item of the data domain normal equation can be determined jointly, thus higher computing efficiency can be attained. Simulation results in smart antenna systems are provided and the performance of QRD - LS algorithm is demonstrated.