Collaborative Beamforming Via Diffusion Adaptation Based On Tensor Over Array Networks

Conventional distributed collaborative beamforming would generally wrestle with the challenges of slowed-down convergence rate, prohibitively high computational costs and inefficient communication when applied to large-scale networks equipped with massive arrays. To address these challenges, we herein reformulate the distributed collaborative beamforming problem from the tensor perspective. By exploiting the inherent algebraic structure present in the problem, we develop a fully distributed collaborative beamforming algorithm incorporating the diffusion scheme for arrays endowed with the property of multi-linear translation invariance (MLTI). We also derive the convergence constraint of the proposed algorithm. Illustrative simulations validate the superior performance of the proposed algorithm, with notably accelerated convergence rate, reduced computational complexity and enhanced communication efficiency. (C) 2020 Elsevier Inc. All rights reserved.