Downlink beamforming design for mobile users in massive MIMO system.

With numerous antennas, the massive multiple-input multiple-output (MIMO) system can precisely point to the user equipments (UEs) with narrow beams. Mobile UEs, however, may suffer from the narrow beam nature of the massive MIMO system, since UEs may move out of the beam coverage. Frequent CSI feedback and beam adjustment help to track mobile UEs. However, a significant amount of transmission and computation overhead is incurred. In this paper, we consider a mobility adaption problem in massive MIMO systems with infrequent CSI feedback. We propose a beam design method to use one beam to serve multiple transmissions, thus reducing the overhead of frequency beam adjustment. An optimization problem is formulated that maximizes the overall system sum rate during one beam adjustment, i.e., during the entire time interval between CSI feedback. We assume that each transmission antenna has a peak power constraint, as the power amplifier in each branch limits the transmit power. A fractional programming based algorithm is proposed to solve the problem. To further reduce the computation overhead, a sub-optimal solution is obtained. Simulation results show that the proposed algorithms achieve satisfactory performance with little transmission and computation overhead compared to existing algorithms. [ABSTRACT FROM AUTHOR]