Non-orthogonal Multiple Access Based Hybrid Beamforming in 5G mmWave Systems

In this paper, we propose a non-orthogonal multiple access(NOMA) based hybrid beamforming design in 5G mmWave systems. The proposed design depends on the known array geometry and incurs a low training and feedback overhead. Our model assumes that each user employs analog-only beamforming while the BS performs hybrid analog and digital beamforming. To achieve a higher sum capacity, multiple users are allowed to share the same beam based on NOMA mechanism. However, the use of analog beamformer from different localized users means that the BS's digital beamformers are not perfectly aligned with the users' baseband effective channels and multiple users may be assigned similar or even identical beamformers. In the meantime, a user pairing and power allocation algorithm is proposed to mitigate interference from other beams so as to maximize the sum capacity. Simulation results verify the significant advantage of the proposed scheme in sum capacity.