Hybrid Beamforming for Massive MIMO Over-the-Air Computation

Over-the-air computation (AirComp) has been recognized as a promising technique in Internet-of-Things (IoT) networks for fast data aggregation from large number of wireless devices. However, the accuracy AirComp highly depends on devices with worst channels condition, which degrades severely when becomes large. To address this issue, we exploit massive multiple-input multiple-output (MIMO) hybrid beamforming, order to enhance computational cost-effective manner. In particular, consider scenario multi-antenna simultaneously sending an access point (AP) equipped antennas functional over air. Under setup, jointly optimize transmit digital beamforming at and receive AP, objective minimizing mean-squared error (MSE) subject individual power constraints solve non-convex design optimization problem, propose alternating-optimization-based approach, analog are optimized alternating two computationally efficient algorithms handle challenging by exploiting techniques successive convex approximation (SCA) coordinate descent (CD), respectively. It is shown that special case fully-digital receiver achieved MSE MIMO system inversely proportional antennas. Furthermore, numerical results show proposed substantially enhances performance compared other benchmark schemes, while SCA-based algorithm performs closely upper bound beamforming.