Extremely Large-scale Lens Antenna Array: Location-aware Near-field Beamforming

Previous works on near-field communications have mostly considered the extremely large-scale array (XL-array), for which each antenna is attached with a phase shifter (PS), hence resulting in high hardware and power cost. To address this issue, we consider in this paper the more cost-effective extremely large-scale lens antenna array (XL-LAA), where only a few PSs are deployed and connected to selected lens antennas for receiving spread signal powers under the near-field channel model. Based on this hardware architecture, we formulate an optimization problem to maximize the system (cost) efficiency under random user location, by jointly optimizing the number of PSs to be deployed, as well as the switch network and phase shifts. Although this problem is non-convex and hence hard to solve, we propose a new and efficient algorithm that firstly designs a location-aware XL-LAA receive beamforming method given any user location, and then determines the optimal number of PSs deployed based on the user location distribution. Finally, numerical results are presented to validate the effectiveness of proposed XL-LAA architecture as well as the superior system efficiency of the proposed algorithm as compared to various baseline schemes.