Fully -/Partially-Connected Hybrid Beamforming for Multiuser mmWave MIMO Systems

Hybrid analog and digital beamforming (HBF) can be utilized to reduce the number of radio frequency chains in millimeter-wave massive multiple-input multiple-output (MIMO) systems. This letter investigates a HBF design for multiuser MIMO systems. By applying the fractional programming, an universal algorithm is proposed to maximize the sum rate of the multiuser system with the fully-connected and partially-connected architectures. The closed-form digital beamformers are updated by the prox-linear block coordinate descent (BCD) method. Then we transform the non-convex analog precoder optimization problem into a quadratically constrained quadratic program (QCQP), which is widely existed in analog beamforming and reconfigurable intelligent surface design. An iterative coordinate ascent (ICA) algorithm is proposed for the kind of problem, which can obtain high-performance solutions and much more computationally efficient than the methods using linear approximations. Moreover, we give an optimality condition to ensure the converged solution is globally optimal. Numerical results show the superior performance of the proposed ICA algorithm and HBF scheme.