应用科学学报 ›› 2017, Vol. 35 ›› Issue (2): 207-216.doi: 10.3969/j.issn.0255-8297.2017.02.007

• 信号与信息处理 • 上一篇    下一篇

自适应变步长并行Laguerre盲均衡算法研究

崔琳1,2, 杨俊安1,2, 王伟1,2, 刘辉1,2   

  1. 1. 电子工程学院, 合肥 230037;
    2. 安徽省电子制约技术重点实验室, 合肥 230037
  • 收稿日期:2016-05-12 修回日期:2016-11-14 出版日期:2017-03-30 发布日期:2017-03-30
  • 通信作者: 杨俊安,教授,博导,研究方向:信号处理、智能计算等,E-mail:yangjunan@ustc.edu E-mail:yangjunan@ustc.edu
  • 基金资助:

    安徽省自然科学基金(No.1308085QF99)资助

Blind Equalization with Adaptive Variable Step Sizes Based on Parallel Laguerre Filter

CUI Lin1,2, YANG Jun-an1,2, WANG Wei1,2, LIU Hui1,2   

  1. 1. Electronic Engineering Institute, Hefei 230037, China;
    2. Key Laboratory of Electronic Restricting Technique of Anhui Province, Hefei 230037, China
  • Received:2016-05-12 Revised:2016-11-14 Online:2017-03-30 Published:2017-03-30

摘要:

深衰落信道会引起抽头系数长和误码率高等问题.在分析现有滤波器结构及盲均衡算法的基础上,提出了自适应变步长并行Laguerre盲均衡算法.该算法采用Laguerre序列代替传统抽头延迟,缩短了抽头系数,并借助判决引导算法和分数间隔思想,加快了算法收敛速度.鉴于步长因子对盲均衡器收敛速度、剩余码间干扰影响较大,对不同自适应步长算法进行了分析和比较.在此基础上提出了适用于高速通用数据链信号盲均衡的步长修正方案.实验表明该算法较传统算法在收敛速度、误码性能方面均有所提升.

关键词: 盲均衡, Laguerre滤波器, 通用数据链, 变步长

Abstract:

A blind equalization algorithm with adaptive variable step sizes based on parallel Laguerre filter is proposed to overcome problems such as long tap coefficients, and high bit error rate (BER) caused by deep fading channels. By analyzing existing filter structures and blind equalization algorithms, the proposed algorithm replaces traditional tapped delay with Laguerre sequence to make tap coefficients shorter. Meanwhile, convergence is accelerated by using a decision-directed algorithm and a fractionally spaced structure. Since the step size has a great effect on convergence speed and residual error, different algorithms with variable step sizes are compared. Based on the comparison, a revised equalization algorithm structure is devised to achieve the most suitable step sizes for high-speed common data link (CDL). Compared with traditional blind equalization algorithms, experimental results show that the proposed method improves equalization performance and reduces BERs remarkably.

Key words: blind equalization, variable step size, common data link, Laguerre filter

中图分类号: