摘要
该文描述了在Matlab中编程实现语音通信中去除噪声技术。依据输入信号在迭代过程中估计梯度矢量、更新权系数以达到最优的自适应迭代算法,采用一种期望响应和滤波输出信号之间误差的均方值最小(LMS算法)为准则的梯度最陡下降方法。讨论收敛因子μ的取值范围使降噪效果达到最优。
This article describes to remove noise in Matlab programming realization of voice communication technology.On the basis of the input signal in the process of iteration estimates gradient vector,update the weights in order to achieve the optimal adaptive iteration algorithm.Using an expected response and filtering the mean square error between the output singals as the criterion of minimum gradient steepest descent method.Discuss the scope of convergence factor causes the noise reduction effect to achieve the optimal.
出处
《电脑知识与技术（过刊）》
2014年第11X期7650-7652,共3页
Computer Knowledge and Technology
关键词
最小均方LMS算法
期望值
最陡下降法
权系数向量
收敛因子
Minimum mean square LMS algorithm
expected value
steepest descent method
Weight vector
convergence factor
