The Research On Feedback Cancellation Algorithm Of Digital Hearing Aids

Hearing impaired patients can improve their hearing level by wearing digital hearing aids.However,echo feedback seriously affects the usage of the digital hearing aid.This paper focuses on the technology of traditional echo feedback cancellation and also proposes the improved algorithms.The main contents of the thesis include:Firstly,this paper studies the principle of the generation of speech and hearing.Several digital models of speech signal generation and common methods of speech signal research are introduced.Several important techniques in digital hearing aids are also introduced,including echo feedback cancellation,speech enhancement,sound source localization and loudness compensation.The echo feedback cancellation is mainly introduced.Secondly,the block updating NLMS algorithm is proposed.This paper uses batch update on the update frequency of the filter coefficients and introduces the idea of variable step-size in the updating of the iterative step.The simulation shows that the improved algorithm called INBLMS has the faster speed of convergence and lower steady-state offset.Finally,this paper chooses the affine projection algorithm.In order to solve the contradiction between steady-state offset and convergence speed,this paper introduces an idea of variable step-size in APA.In this paper,a multi-forgetting factor is proposed based on the variable step-size affine projection algorithm,which can be used to speed up the convergence rate and reduce the steady-state error.In order to reduce the computational complexity,this paper also introduces the idea of variable projection order.The simulation shows that the improved algorithm can achieve good results in terms of convergence speed,steady-state error and computational complexity whether the input signal is a Gaussian random signal or a real speech signal.