Beam Training for Multiuser XL-MIMO Systems: A Graph Neural Network Approach

Extremely large-scale multiple-input multiple-output (XL-MIMO) is regarded as one of the key technologies for future 6G networks, which can further improve spectral efficiency by deploying far more antennas than conventional massive MIMO systems. However, beam training in multiuser XL-MIMO systems is challenging. To tackle this issue, we propose a graph neural network (GNN)-based beam training scheme for the multiuser XL-MIMO system, in which only the far-field wide beams need to be tested for each user. Specifically, the GNN is utilized to map the beamforming gain information of the far-field wide beams to the optimal near-field beam for each user, where the information of the surrounding users can also be utilized by the GNN to further improve the accuracy of the beam training. Simulation results show that the performance of the proposed scheme can approach that of the exhaustive scheme but has more than a 93 % reduction in the pilot overhead.