12IoT-Based Bias Analysis in Acoustic Feedback Using Time-Variant Adaptive Algorithm in Hearing Aids
G. Jayanthi1* and Latha Parthiban2
1Sathyabama Institute of Science and Technology, Department of Electronics and Communication Engineering, Chennai, India
2Pondicherry University CC, Department of Computer Science, Pondicherry, India
Abstract
A feedback signal (FS) is generated in hearing aids when the loudspeaker signal (LS) has superimposed on the input sound signal, which is possibly contaminated by the background noise. This feedback will emulate an ambient linkage between the loudspeaker and microphone signals that generate irritating whistling sounds to the user, during which the gain of the system gets reduced.
The adaptive filters have been employed in acoustic feedback cancellation, for estimating the feedback response in hearing aids using the adaptive filter algorithms. However, bias is introduced into the system, due to the spectrally colored nature of the input signal. Several approaches have been proposed to cancel this bias using algorithms based on continuous and non-continuous adaptation methods, Prediction Error Method (PEM), Wavelet Transform method, etc.
The general key to using an adaptive Least-Mean-Square (LMS) algorithm is to update the filter coefficients individually and independently of each other. The prime parameter, the step-size is used for updating the filter coefficients, for which it is kept fixed throughout the process. This will affect the ...
Get Human Communication Technology now with the O’Reilly learning platform.
O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.
