A Universal Linear Beamformer for Microphone Arrays

Existing beamforming algorithms often lack sufficient flexibility in balancing multiple objectives. This paper presents a universal linear beamformer for microphone array applications, built upon a highly general optimization criterion that incorporates several design parameters. These parameters enable a flexible tradeoff among desired signal distortion, interference suppression, and noise reduction. Based on this criterion, we derive a universal beamformer and give several equivalent formulations. We show that the proposed beamformer encompasses most of the known and classical adaptive beamformers. Additionally, with the help of simple noise models, we propose a two-step approach to the development of a wide variety of fixed beamformers tailored to different applications. The effectiveness of our proposed beamformer on tradeoff among different performance measures is demonstrated through simulation evaluations.