An Efficient Random Beamforming Method based on Beam Subset Selection

In the finite number of users, the proposed beam subset selection method improves the throughput performance of the orthogonal random beamforming (ORBF), which is a scheme to achieve asymptotic sum capacity for a downlink multiuser multiple-input/multiple output (MIMO) system using partial channel information in the infinite number of users. We show that there exists an optimal number of random beams to maximize the sum throughput for given number of users. When the number of random beams is less than the number of antennas, selecting the random beams to maximize signal- to-interference-plus-noise ratios (SINRs) provides additional SINR gain to ORBF. Through the simulation results, we show the performance improvement by the proposed beam subset selection.