基于子空间新型跟踪策略的RLS盲多用户检测算法

摘要

To overcome the issue that the performance of the subspace-based blind multiuser detector degrades due to underestimation of the sources' number in direct sequence/code division multiple access (DS/CDMA)systems, a novel subspace tracking strategy is presented and the recursive least square (RI.S) blind multiuser detection algorithm is modified. As for the novel subspace tracking strategy, the desired user's signature waveform constitutes one part of the signal subspace and its orthogonal complement is tracked as the other parts of that signal. This brings the benefits of a small signal subspace estimation error and lower computational complexity. Even in the underestimation of the sources' number, the performance of the detector does not degrade significantly because the signal subspace always maintains the information on the desired user. The RLS blind multiuser detection algorithm is combined with the subspace approach. As a result of forward error feedback, this algorithm can reduce sensitivity to signal subspace estimation. Simulations are provided to demonstrate the effectiveness of the proposed detector.%针对DS/CDMA系统中信源个数欠估计时基于子空间的盲多用户检测算法性能恶化的问题,提出了子空间新型跟踪策略,并对递归最小二乘(RLS)盲多用户检测算法进行改进.在新型跟踪策略中,目标用户特征波形作为信号子空间的一部分,仅跟踪信号子空间中目标用户特征波形的正交分量,然后构造信号子空间,从而提高了信号子空间准确度,且计算复杂度较低,即使在信源个数出现欠估计时,目标用户信息也未丢失,检测性能所受影响较小;将子空间引入到RLS盲多用户检测算法中,有效降低了对信号子空间跟踪准确度的要求.仿真结果表明了该算法的有效性.