Iterative best beam selection for random unitary beamforming

Due to multi-user diversity, the performance of a random unitary beamforming (RUB) can be improved upon in proportion to the number of users at the expense of an increase in feedback. It is essential to identify a mechanism to determine a set of orthonormal beams for improving the channel capacity. This paper proposes a unique scheme, named iterative best beam selection-based RUB (I-RUB), in which the base station selects the best beamforming vectors from unitary matrices generated by the proposed iterative algorithm. Using the selected beamforming vector at each stage, the beam set can ultimately be constructed for high sum-rate capacity of downlink, which leads to a reduction in feedback compared to those of other RUB schemes based on the beam set selection.