|本期目录/Table of Contents|

[1]余尤好.神经网络在通信系统回音对消中的应用[J].武汉工程大学学报,2012,(9):70-74.[doi:103969/jissn16742869201209016]
 YU You hao.Application of neural network in echo cancellation of communication system[J].Journal of Wuhan Institute of Technology,2012,(9):70-74.[doi:103969/jissn16742869201209016]
点击复制


神经网络在通信系统回音对消中的应用
(/HTML)
分享到:

《武汉工程大学学报》[ISSN:1674-2869/CN:42-1779/TQ]

卷:
期数:
2012年9期
页码:
70-74
栏目:
机电与信息工程
出版日期:
2012-10-10

文章信息/Info

Title:
Application of neural network in echo cancellation of communication system
文章编号:
16742869(2012)09007005
作者:
余尤好
莆田学院电子信息工程系,福建 莆田 351100
Author(s):
YU Youhao
Department of Electronics and Information Engineering, Putian University, Putian 351100, China
关键词:
回音对消滤波器神经网络语音通话
Keywords:
echo cancellation filter neural network speech conversation
分类号:
TN912.3
DOI:
103969/jissn16742869201209016
文献标志码:
A
摘要:
通信系统中的回音影响通话质量,造成通话困难.为控制通信系统回音的影响,利用自适应线性神经网络建立回音对消系统,求出话音经过传输过程发生非线性变化后的结果.对神经网络学习速率和训练次数进行调节,获得回音信号的最佳估计值,原始信号与之相减实现回音对消.实验时分别对通信系统直接回音和间接回音进行消除,并利用MATLAB软件绘出对应信号的频谱图和语谱图,以便观察分析.仿真结果表明,消除回音后有用的语音信号得到增强,回音被有效削弱,通话具有较高的信噪比,人耳感受基本没有回音.
Abstract:
Echo of the system affects the speech quality seriously and makes it difficult to communicate. Adaptive linear neural network was used to control echo effects of communication system. The echo cancellation system was established by adaptive linear neural network, and nonlinear variation results of voice was obtained. Adjusting the learning rate and the number of training, the best estimate of the echo signal was obtained by adaptive linear neural network, subtracting it from the original signal, and then echo cancellation was realized. Direct echo and indirect echo were eliminated from communication system in the experiments. In order to observe and analyze the results, MATLAB software was used to draw the corresponding signal spectrum and spectrogram. The simulation results show that the signal is enhanced after echo cancellation and echo is effectively reduced; the call has a high signaltonoise ratio and people feel no echo.

参考文献/References:

[1]周开利,康耀红. 神经网络模型及其MATLAB仿真程序设计\[M\]. 北京:清华大学出版社,2005.
[2]郭峰,任兴民,刘婷婷. 基于神经网络消噪的独立成分分析方法研究\[J\]. 机械科学与技术,2010,29(12):16781682.
[3]张帆. Adaline神经网络随机逼近LMS算法的仿真研究\[J\]. 电子设计工程,2009,17(9):8890.
[4]董翠英,周长英. 基于RBF神经网络的语音信号的噪声消除\[J\]. 制造业自动化,2010,32(5):228230.
[5]高宁,郑恩让,马令坤,等. 基于神经元网络的自适应噪声抵消系统研究与实现\[J\].计算机测量与控制,2010,18(1):133135.
[6]李政洋. 基于AD神经网络的语音增强\[D\]. 苏州:苏州大学,2008.

相似文献/References:

[1]赵振华,梁凯.基于线阵CCD的IIR算法设计与实现[J].武汉工程大学学报,2008,(03):97.
 ZHAO Zhen hua,LIANG Kai.Design and implementation of IIR algorithms based on linear CCD[J].Journal of Wuhan Institute of Technology,2008,(9):97.

备注/Memo

备注/Memo:
收稿日期:20120813基金项目:福建省教育科学“十二五”规划2012年度常规课题(FJCGGJ12034)作者简介:余尤好(1977),男,福建莆田人,讲师,硕士.研究方向:数字信号处理.
更新日期/Last Update: