Research Of Beamforming Algorithm In Digital Hearing Aids

Population aging will become a major character in the population structure in the future, which means that the number of elderly people will increase greatly. Hearing loss is one of the most common health problems in aged population. Most aged people have the hearing loss problem in different extent. We can predict that hearing aid market in China will have a enormous potentiality and a wide development space in the future.Denoising algorithm is one of the most important algorithms in hearing aids, but the traditional denoising algorithms can not simulate the "cocktail party effect" of human ear. To solve this problem, this paper studied the beamforming algorithm.After a thorough understanding of the beamforming algorithm, we derived and implemented the frequently-used beamforming algorithms, including the first-order differential microphone structure(FDM), the minimum variance distortion free response(MVDR), and the generalized sidelobe cancellation(GSC) algorithm.Since the binaural hearing aid can achieve better result than that of monaural hearing aids in noise suppression, acoustic source localization and speech intelligibility improving, In this paper, the BMVDR algorithm used in binaural hearing aids is proposed according to the structural characteristics of binaural hearing aids. When the number of the microphones in binaural hearing aid is twice than monaural hearing aid,the output in each side of BMVDR will process the signal by utilizing a linear filter whose number of the filter coefficient is twice than monaural hearing aids. So BMVDR has a better performance than FDM and MVDR in noise reduction. The BMVDR is used to process the signal received by all the microphones in binaural hearing aids,which makes the amount of the array element more than that of monaural hearing aids.This makes BMVDR have better spatial resolution than FDM and MVDR. After deducing, analyzing and simulating the BMVDR, the results of theoretical analysis are verified by the simulation results.Compared with monaural hearing aid, a significant advantage of binaural hearing aid is that it can preserve the binaural cues. In this paper, we discussed the retention ability of the BMVDR for the interaural time difference(ITD). Because the BMVDR algorithm can suppress the noise signal in the maximum extent, it does not have the ability to retain the spatial information of the directional noise.The BMVDR-? and BMVDR-ITF are designed to solve this problem. The BMVDR-?can effectively retain the spatial information of the noise from a certain direction by adding the weight factor ?. The BMVDR-ITF realizes the reservation of the spatial information of the directional noise by adding the restricted condition that the input and output ITF of directional noise is equal. Since the constraint conditions of BMVDR-ITFare less than BMVDR-?, the BMVDR-ITF has a better performance in noise reduction.However, the reservation of spatial information of the directional noise is at the cost of reducing output SNR. Finally, the algorithm is simulated on PC and the results verified the theoretical analysis.