基于嵌入式的助听器回波算法研究与实现

Research and Implementation of Echo Cancellation in Hearing Aids Based on Embedded System

回声会导致助听器产生啸叫,损坏助听器设备,破坏患者的残余听力。为此,本文在助听器回声抵消模型的基础上,针对输入信号的能量变化,研究了基于归一化最小均方NLMS(Normalized Least Mean Squares)的自适应助听器回声抵消算法。通过对比LMS和NLMS两种算法的MSE,ERLE等性能,研究发现NLMS算法在数字助听器模型中有更好的回声抵消性能。此外,将NLMS算法移植到嵌入式平台中,并通过实验对比算法的性能。因此,本研究对于助听器的回声抵消算法设计具有很强的实用价值。

Echo leads to howling in hearing aids,demolishes the hearing device and impairs the residual hearing of the patients.Adaptive echo cancellation in hearing aids based on normalized least mean squares（NLMS）algorithm was studied.By comparing the MSE,ERLE of LMS algorithm and NLMS algorithm,we found that NLMS algorithm has more preferable echo cancellation performance in digital hearing aids.In addition,NLMS algorithm was implanted into embedded platform.Experiments validate the performance of the NLMS algorithm.So it is of important practice value for echo cancellation in hearing aids.