Research On Echo Cancellation Technology Of Hearing Aid System Based On Gammatone Filter

As a medical device,digital hearing aid is an effective way to compensate for hearing loss.With the gradual development of digital signal processing,the volume of digital chips in digital hearing aids is decreasing,which makes the overall volume of digital hearing aids smaller and more convenient to wear.Digital hearing aids can be adjusted according to the personal situation of hearing loss patients,so this is also the reason why patients with hearing impairment choose digital hearing aids.Because the sound output from the loudspeaker in the digital hearing aid will be picked up by the microphone again and produce echo,the serious echo will directly affect the use experience of hearing impaired patients.Therefore,this paper mainly studies the echo cancellation technology in digital hearing aid system.Aiming at the slow convergence speed and large steady state of normalized least mean square algorithm,affine projection algorithm,proportional normalized least mean square algorithm and particle swarm optimization algorithm,this paper proposes a new high-speed adaptive filtering method for hearing aid system to achieve effective echo cancellation.Firstly,this paper studies the echo model of conventional digital hearing aid system.Through the research and analysis of the model,it is found that the echo signal and the original speech signal are highly correlated.Therefore,this paper studies several existing adaptive echo cancellation methods,including normalized least mean square algorithm,affine projection algorithm,proportional normalized least mean square algorithm and particle swarm optimization algorithm Through comparative analysis,it is found that the first three methods have the problems of slow convergence speed and large steady-state error.PSO method has fast convergence speed,but steady-state error is not significantly reduced.In order to effectively improve the shortcomings of PSO algorithm,this paper introduces gamma tone filter on the basis of PSO algorithm,and uses Gammatone filter to process frequency band to reduce the correlation of input echo signal.Meanwhile,it fully reflects the difference of information in each frequency band and fully retains the weight in different frequency bands You need voice information.After the Gammatone filter is used to divide the frequency band,the particle swarm optimization algorithm is used to de echo each frequency band.Finally,the pure speech of multiple subbands is reconstructed.The simulation results show that the proposed particle swarm optimization echo cancellation algorithm based on gamma tone filter has lower steady-state error than the traditional adaptive echo cancellation algorithm,and the echo cancellation effect of the hearing aid system is obviously better than other methods.