首页 | 本学科首页   官方微博 | 高级检索  
     

一种改进的快速LMS/Newton算法及其在声回波对消中的应用
引用本文:王永德,王春霞. 一种改进的快速LMS/Newton算法及其在声回波对消中的应用[J]. 电子学报, 2000, 28(2): 97-100
作者姓名:王永德  王春霞
作者单位:四川大学无线电系,四川成都 610064
基金项目:国家自然科学基金!(No .69672 0 39)
摘    要:本文提出了一种改进的快速LMS/Newton算法.该算法既不同于完全基于RLS算法的FNTF算法[1],又不同于完全基于LMS算法的算法2[2].算法在格型预测器部分采用快速RLS算法,而在自适应横向滤波器部分采用LMS算法,并加以适当改进.所提出的算法既不存在FNTF的稳定性问题,又比算法2[2]有更快的收敛速度.采用语音信号激励,辨识实际测量的声回授通道的冲激响应的模拟实验证实了上述论断.

关 键 词:LMS/Newton算法  自适应系统辨识  声回波对消  
收稿时间:1998-11-13

An Improved Fast LMS/Newton Algorithm for Application to Acoustic Echo Cancellation
WANG Yong-de,WANG Chun-xia. An Improved Fast LMS/Newton Algorithm for Application to Acoustic Echo Cancellation[J]. Acta Electronica Sinica, 2000, 28(2): 97-100
Authors:WANG Yong-de  WANG Chun-xia
Affiliation:Dept.of Radio Electronics,Sichuan University,Chengdu 610064,China
Abstract:In this paper we proposed an improved fast LMS/Newton algorithm for application to acoustic echo cancellation.The new algorithm is different from not only the FNTF algorithm [1] based the FRLS,but also the algorithm 2 [2] based on the LMS completely.For the proposed one we use the fast RLS algorithm in lattice predictive filtering part with some modification and the LMS in adaptive transversal filtering part.Therefore it has the advantage of both the fast convergence for FNTF and the stability for the algorithm 2 [2] .The simulation experiments using Chinese speech exciting for modeling actual acoustic channel justify the previous issue.
Keywords:LMS/Newton algorithm  adaptive system identification  acoustic echo cancellation
本文献已被 CNKI 维普 万方数据 等数据库收录!
点击此处可从《电子学报》浏览原始摘要信息
点击此处可从《电子学报》下载免费的PDF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