东北大学学报（自然科学版）

东北大学学报:自然科学版 ›› 2014, Vol. 35 ›› Issue (9): 1253-1257.DOI: 10.12068/j.issn.1005-3026.2014.09.009

• 信息与控制 • 上一篇    下一篇

基于复数FastICA算法的盲多用户检测

刘晓志1，邵秀1,2，杨英华1，陈晓波1   

  1. (1 东北大学 信息科学与工程学院， 辽宁 沈阳110819； 2 沈阳辽海装备有限责任公司， 辽宁 沈阳110003)
  • 收稿日期:2013-05-06 修回日期:2013-05-06 出版日期:2014-09-15 发布日期:2014-04-11
  • 通讯作者: 刘晓志
  • 作者简介:刘晓志(1968-)，女，辽宁沈阳人，东北大学副教授；陈晓波(1960-)，男，辽宁沈阳人，东北大学教授.
  • 基金资助:
    中央高校基本科研业务费专项资金资助项目(N100404018).

Blind Multiuser Detection Based on Complex FastICA Algorithm

LIU Xiaozhi1， SHAO Xiu1,2，YANG Yinghua1， CHEN Xiaobo1   

  1. 1 School of Information Science & Engineering， Northeastern University， Shenyang 110819， China; 2 Shenyang Liaohai Equipment Co.， Ltd.， Shenyang 110003， China.
  • Received:2013-05-06 Revised:2013-05-06 Online:2014-09-15 Published:2014-04-11
  • Contact: LIU Xiaozhi
  • About author:-
  • Supported by:
    -

RichHTML

1

PDF (PC)

1305

摘要： 针对MIMO-OFDM系统，提出了一种基于复数FastICA的盲多用户检测算法.该算法首先利用复数FastICA算法的快速收敛特性来提高多用户的分离速度，同时利用信号的相关函数对复数FastICA算法引起的幅度不确定性和相位不确定性进行了修正，最后将所提出的改进算法与复数自然梯度学习算法(CNGLA)进行仿真比较.结果表明，相比于传统的自然梯度算法，所提算法不仅收敛速度较快，而且具有更低的误码率，另外随着接收天线的增多，信号的分离效果会更好.

关键词: MIMO-OFDM系统, 复数FastICA算法, 盲多用户检测, 幅度不确定性, 相位不确定性

Abstract: A new blind multiuser detection algorithm was proposed on the basis of complex FastICA in MIMOOFDM system. With the proposed algorithm， the separation of multiuser could be speeded up by fast convergence characteristics of complex FastICA algorithm firstly， and the indeterminacy of amplitude and indeterminacy of phase by CFastICA algorithm were revised by using correlation function. At last， comparing the improved algorithm with the complex natural gradient learning algorithm(CNGLA)， it was proved that CFastICA algorithm has not only faster convergent speed but also smaller bit error rate. It also indicated that using more numbers of receiving antennas would get better signal separation.

Key words: MIMOOFDM system, complex FastICA algorithm, blind multiuser detection, indeterminacy of amplitude, indeterminacy of phase

中图分类号: