Beamforming with Intelligent Metasurfaces: Operating Principles and Possible Implementations

This article discusses a beamforming architecture based on the Programmable Metasurfaces (PMS) for use in future telecommunications. Reflecting beamforming PMS operate by controlling the local reflection phase while omitting the amplitude information. By optimizing the reflection phase distribution using different algorithms, the sidelobe levels are minimized. The acquired phase distributions obtained from the optimization simulations can be used to train a neural network to realize dynamic adaptive beamforming in order to track and control communication beams with predefined characteristics. Here, we discuss the ongoing work towards this goal and present related mathematical models and some simulations results.