Multibeam Tracking for Future Multi-User Communications Employing Generalized Joined Coupler Matrix
Charles A. Guo, Qingqing Cheng, Jinhong Yuan · 2024
Effective multiuser communications require multibeam antennas whose beams can be tracked in real time to maintain link quality. Whilst there is a lot of literature tackling this issue for digital beamformers and MIMO, it remains a challenge for analogue beamformers using radio frequency (RF) circuits. For next-generation wireless communication networks, it is necessary to provide coverage for dynamic users with minimal energy costs, and analogue multibeam beamformers serve as an attractive candidate. In this paper, we propose a novel framework for achieving multibeam tracking of analogue multibeam antennas using a combination of the recently developed generalized joined coupler (GJC) matrix and the DDPG deep reinforcement learning method. The algorithm we introduced enables an automated agent to control the parameters of a GJC matrix in response to the direction changes of user signals to maximize the beamforming gain whilst simultaneously enabling angle-of-arrival estimation. The proposed approach is verified by simulations.