Convex Optimization for Distributed Acoustic Beamforming

Beamforming filter optimization can be performed over a distributed wireless sensor network, but the output calculation remains either centralized or linked in time to the weights optimization. We propose a distributed method for calculating the beamformer output which is independent of the filter optimization. The new method trades a small decrease in signal to noise performance for a large decrease in transmission power. Background is given on distributed convex optimization and acoustic beamforming. The new model is described with analysis of its behaviour under independent noise. Simulation results demonstrate the desirable properties of the new model in comparison with centralized output computation.