首页 | 本学科首页   官方微博 | 高级检索  
     

一种改进的解相关LMS自适应算法
引用本文:段正华,王梓展,鲁薇. 一种改进的解相关LMS自适应算法[J]. 湖南大学学报(自然科学版), 2006, 33(3): 114-118
作者姓名:段正华  王梓展  鲁薇
作者单位:湖南大学,计算机与通信学院,湖南,长沙,410082;湖南大学,计算机与通信学院,湖南,长沙,410082;株洲电业局,湖南,株洲,412000;湖南铁道职业技术学院,湖南,株洲,412000
基金项目:国家自然科学基金资助项目(20040042082)
摘    要:针对变步长LMS自适应滤波算法在输入信号高度相关时,收敛速度下降导致性能下降的问题,提出了一种改进的解相关LMS自适应算法,该算法引入解相关原理和归一化处理,用输入向量的正交分量来更新滤波器权系数,有效加快了算法的收敛速度,且稳态误差小,使得算法在有色输入和大范围的动态输入下都能保持良好性能.

关 键 词:自适应算法  归一化  变步长  LMS(least mean square)
文章编号:1000-2472(2006)03-0114-05
收稿时间:2005-08-09
修稿时间:2005-08-09

A Modified Uncorrelated LMS Adaptive Algorithm
DUAN Zheng-hu,WANG Zi-zhan,LU Wei. A Modified Uncorrelated LMS Adaptive Algorithm[J]. Journal of Hunan University(Naturnal Science), 2006, 33(3): 114-118
Authors:DUAN Zheng-hu  WANG Zi-zhan  LU Wei
Affiliation:1.College of Computer and Communication, Hunan Univ, Changsha,Hunan 410082,China; 2.Zhuzhou Electric Power Supply Bureau, Zhuzhou,Hunan 412000,China; 3.Hunan Railway Professional Technology College,Zhuzhou,Hunan 412000,China
Abstract:We discussed and analyzed the Variable Step-Size Adaptive Algorithm(VSSLMS),in which the descent of its convergence speed led to the reduction of performance when the input signals were heavily correlated.As a result,we proposed a modified uncorrelated LMS adaptive algorithm.We substituted the orthogonal components of the input signals for input vectors to update coefficients of the adaptive filter.Due to the uncorrelation principle and the process of normalization,the convergence speed could be quickened and the misadjustment could be smaller,so the algorithm could achieve good performance even for colored inputs and with a large dynamic input range.
Keywords:adaptive algorithon  normalization  variable step-size  LMS(least mean square)  
本文献已被 CNKI 维普 万方数据 等数据库收录!
点击此处可从《湖南大学学报(自然科学版)》浏览原始摘要信息
点击此处可从《湖南大学学报(自然科学版)》下载全文
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号