Robust adaptive processing of microphone array data for hearing aids

Abstract

The problem of adaptively combining the outputs of an array of microphones as a single input for a hearing aid is investigated. A robust processor based on a constrained minimum variance optimization approach is used. One fundamental criteria employed in designing this robust beamformer limits the amount of cancellation of the desired signal. The results presented include the effects of acoustic headshadow, small room reverberation, microphone placement uncertainty, and desired speaker location uncertainty. Performance improvement is measured as a predicted change in the speech reception threshold (SRT) between single microphone and multi-microphone conditions. Performance improvements are demonstrated relative to the "best" single microphone in the array for block optimum and adaptive spatial filters. The performance of the block optimum arrays is shown to be attainable with adaptive implementations. A fast-attack, slow release input signal power averager allows the adaptive processor to avoid instabilities commonly experienced with nonstationary, impulsive inputs such as speech.<>