Acoustic Beamforming for Hearing Aids Using Multi Microphone Array by Designing Graphical User Interface

Abstract: Hearing impaired persons lose their ability to distinguish speech signal in ambient noise. Human hearing system is sensitive to interfering noise. Interfering noise decreases the quality and intelligibility of the speech signal which in turn makes speech communication default. To make the speech signal effective and useful for hearing impaired, they need to be enhanced from noisy speech signal. Speech enhancement is one of the most emerging and useful branch in signal processing, to reduce the noise and improves the perceptual quality and intelligibility of the speech signal. Several signal processing techniques has been widely used in hearing aids to enhance the speech signal from the noisy environment. Microphone array is one of the signal processing technique implemented in hearing aids to provide a better solution to the problem encountered by the hearing impaired person when listening to speech in the presence of background noise. Generalized Sidelobe Canceller (GSC) is a powerful technique to enhance the signal of interest which suppressing the interference signal and noise at the output of the array microphones. The main focus of the thesis is to implement a GSC using microphone array, the blocking matrix in the GSC is replaced with Elko’s algorithm. Elko’s algorithm is used to track and attenuate interference or background noise located in the back half plane of the array of microphones. The proposed system is implemented successfully and validated effectively. Clean speech signal is corrupted by various background noises respectively multi-talker babble noise, wind noise, car interior noise, destroyer engine room noise, tank noise, interference male and female voices at five different Signal-to-Noise Ratio (SNR) levels 0db, 5db, 10db, 15db and 20db. Different types of objective tests, such as SNR, Signal-to-Noise Ratio Improvement (SNRI), Perceptual Evaluation of Speech Quality (PESQ), Speech Distortion (SD) and Noise Distortion (ND) are performed on the test set. The platform is made in Matlab Graphical User Interface (GUI) and all the results have been shown by plots produced from Matlab code.