發(fā)布時間：2018-04-17 21:43

  本文選題：數(shù)字助聽器 + 回聲抵消　； 參考：《電子器件》2017年02期

【摘要】：回聲會導致助聽器產(chǎn)生嘯叫,損壞助聽器設備,破壞患者的殘余聽力。為此,本文在助聽器回聲抵消模型的基礎上,針對輸入信號的能量變化,研究了基于歸一化最小均方NLMS(Normalized Least Mean Squares)的自適應助聽器回聲抵消算法。通過對比LMS和NLMS兩種算法的MSE,ERLE等性能,研究發(fā)現(xiàn)NLMS算法在數(shù)字助聽器模型中有更好的回聲抵消性能。此外,將NLMS算法移植到嵌入式平臺中,并通過實驗對比算法的性能。因此,本研究對于助聽器的回聲抵消算法設計具有很強的實用價值。
[Abstract]:Echoes cause hearing aids to howl, damage hearing aids and damage patients' residual hearing.In this paper, based on the echo cancellation model of hearing aids, an adaptive echo cancellation algorithm based on normalized least mean square NLMS(Normalized Least Mean squares is proposed for the energy variation of input signals.By comparing the performance of LMS and NLMS, it is found that the NLMS algorithm has better echo cancellation performance in the digital hearing aid model.In addition, the NLMS algorithm is transplanted to the embedded platform, and the performance of the algorithm is compared through experiments.Therefore, this study has a strong practical value for the design of echo cancellation algorithm for hearing aids.
【作者單位】： 華僑大學工學院;東南大學水聲信號處理教育部重點實驗室;
【基金】：國家自然科學基金項目(61301219,61375028,61571106) 福建省自然科學基金項目(2015J01260) 泉州市科技計劃項目(2014Z103,2014Z113)
【分類號】：R764.5;TN713
，

本文編號：1765384