计算机工程与应用 ›› 2013, Vol. 49 ›› Issue (24): 210-213.

• 信号处理 • 上一篇    下一篇

消除噪声的ICA盲多用户检测

白夏红,李  辉,邢  钢   

  1. 西北工业大学 电子信息学院,西安 710129
  • 出版日期:2013-12-15 发布日期:2013-12-11

Independent Component Analysis based denoising for blind multiuser detector

BAI Xiahong, LI Hui, XING Gang   

  1. School of Electronics & Information, Northwestern Polytechnical University, Xi’an 710129, China
  • Online:2013-12-15 Published:2013-12-11

摘要: 针对已有的基于独立分量分析(Independent Component Analysis,ICA)的盲多用户检测中为了便于分析经常将噪声信号忽略,从而造成系统检测性能下降的问题。提出了基于ICA的盲多用户检测的改进算法,该算法不仅能抑制远近效应和多址干扰,而且利用信号子空间和有效无偏估计原理可最大限度地消除被忽略的噪声,提高了系统的容量。通过Matlab实验仿真,验证了该算法的有效性和实用性。

关键词: 盲多用户检测, 独立分量分析, 直接序列扩频码分多址(DS-CDMA)系统, 罗克拉美下界

Abstract: The noise signal is often neglected for Independent Component Analysis algorithms(ICA) for blind multiuser detection put forward; it leads the detection performance of Direct Sequence spread spectrum Code Division Multiple Address(DS-CDMA) system to decrease. The noise term is taken into account. The improved ICA for blind multiuser detector is proposed. The algorithm can not only suppress Near-Far Effects and Multi-Access Interference but also reduce the noise term neglected as far as possible by signal subspace and effective unbiased estimate, to gain the capacity of the system. Simulation results demonstrate that the proposed method is effective and practicable.

Key words: blind multiuser detection, Independent Component Analysis(ICA), Direct Sequence spread spectrum Code Division Multiple Address(DS-CDMA) systems, Cramer-Rao lower bound