Active Reconfigurable Intelligent Surface-Aided Over-the-Air Computation Networks
Baoyuan Wei, Ping Zhang, Qi Zhang
Abstract
The active reconfigurable intelligent surface (RIS), with the ability to jointly reflect and amplify incident signals, significantly improves the quality of wireless channels. In this letter, we propose an active RIS-aided over-the-air computation (AirComp) network. In the network, an active RIS assists to effectively aggregate the massive data from multiple sensors to the access point. To minimize the computation mean square error, the alternating optimization algorithm is proposed, where the closed-form receiving beamforming at the access point, the semi-closed-form reflection coefficients at the active RIS, and the semi-closed-form transmission scalars at sensors are alternatively updated. Numerical results demonstrate that the proposed active RIS-aided AirComp network outperforms the passive RIS-aided network and the network without RIS.