On implementation of prediction error method-based Adaptive Feedback Cancellation (AFC) in digital hearing aids

Abstract

Acoustic feedback is a notable issue in hearing aid devices, which is produced by the undesired acoustic coupling between the microphone and speaker. Under specific conditions, it creates the howling effect, which is very irritating for the hearing aid user. In this way, the acoustic feedback limits the amplification of hearing aid. The objective of Adaptive Feedback Cancellation (AFC) is to adaptively model the feedback path in such a way that the estimated signal is subtracted from the microphone signal. The main issue in estimating the feedback path model is the correlation between the microphone and speaker signal, which is caused by the closed signal loop. Continuous adaptation feedback cancellation techniques suffer from a modeling error or biasing if the microphone signal is spectrally colored such as speech signals. In this paper, a Prediction Error Method (PEM) based AFC is discussed. The complete system is evaluated extensively for various speech signals for different feedback scenarios with feedback paths measured for real time simulations.