Beamforming with Neural-Networked Programmable Metasurfaces

In this presentation, we outline a novel architecture to enable intelligent, reconfigurable multibeam antennas for future device-to-device wireless communications. Two-dimensional metamaterials - metasurfaces - have been proposed for high-throughput microwave/ millimeter and sub-millimeter band communications. Metasurfaces have emerged as versatile photonic tools for controlling wave fronts and performing nearly-instantaneous operations on the angular spectrum of propagating electromagnetic waves. Our aim is to implement the signal processing and control associated with the formation and tracking of the communication beams within the smart, programmable metasurface layers. Offsetting a part of such processing from the silicon to the quasi-optical metasurface layers shall result in reduction of the computational overheads, which is required when dealing with the evergrowing throughput needs for future telecommunications.