Page 21 - 《应用声学》2022年第6期

第 41 卷 第 6 期 Vol. 41, No. 6
2022 年 11 月 Journal of Applied Acoustics November, 2022

⋄ 研究报告 ⋄

应用于助听器反馈抑制的信噪比自适应滤波算法 ∗

陆悠南 1,2 崔 杰 1† 肖 灵 1

(1 中国科学院声学研究所 北京 100190)

(2 中国科学院大学电子电气与通信工程学院 北京 100049)
摘要：针对基于自适应滤波器的助听器反馈抑制系统，该文提出了一种基于信噪比的归一化最小均方误差算
法，采用最小值统计法估计误差信号的噪声分量，从而计算出误差信号的信噪比来计算自适应滤波系数的更
新步长。当误差信号信噪比越高，语声占主要成分，信号的相关性越强，此时将滤波器的更新步长控制在较小
值，减小滤波器的失调量；当信噪比越低时，噪声占主要成分，信号的相关性相对较弱，更新步长取较大值，加
快滤波器的收敛速度。在仿真实验中，该文提出的基于信噪比的归一化最小均方误差算法相较于传统算法在
平均稳态失调量和稳态失调范围上分别低 1 dB 和 2 dB，其最大稳态增益提高了 4 dB，同时具有更快的稳态
收敛速度，验证了该文提出算法的有效性。
关键词：助听器；反馈抑制；自适应滤波；信噪比
中图法分类号: R764.5 文献标识码: A 文章编号: 1000-310X(2022)06-0867-08
DOI: 10.11684/j.issn.1000-310X.2022.06.003

Adaptive directivity algorithm for hearing aids feedback cancellation based on

signal-to-noise ratio

LU Younan 1,2 CUI Jie 1 XIAO Ling 1

(1 Institute of Acoustics, Chinese Academy of Sciences, Beijing 100190, China)
(2 School of Electronic, Electrical and Communication Engineering, University of Chinese Academy of Sciences,
Beijing 100049, China)

Abstract: Aiming at the hearing aid feedback cancellation system based on the adaptive filter, a normalized
minimum mean-square algorithm based on signal to noise ratio (SNR) is proposed, which uses the minimum
statistical method to calculate the noise component of the error signal and calculate SNR of error signal to
construct the update step size of the adaptive filter coefficient. When the SNR of the error signal is higher,
the voice is the main component. With the correlation of the signal is stronger, the maladjustment of the
adaptive filter is larger. At this time, the update step size of the filter is controlled to a small value to
reduce the maladjustment of the filter. When the SNR is lower, the noise is the main component and the
correlation of the signal is relatively weak. At this time, the update step size takes a larger value to accelerate
the convergence speed of the filter. In the simulation experiment, normalized minimum mean-square algorithm
based on SNR proposed in this paper is 1 dB and 2 dB lower than the traditional algorithm in the mean steady-
state misalignment and the steady-state misalignment range respectively, and the maximumsteady-state gain
is increased by 4 dB, simultaneously the convergence speed is faster than that, which verifies the effectiveness
of the algorithm proposed in this paper.
Keywords: Hearing aids; Feedback cancellation; Adaptive filter; Signal to noise ratio

2021-09-08 收稿; 2021-11-30 定稿
国家重点研发计划项目 (2020YFC2004004)
∗
作者简介: 陆悠南 (1998– ), 女, 江苏宿迁人, 硕士研究生, 研究方向: 信号与信息处理。
† 通信作者 E-mail: cuij@mail.ioa.ac.cn