A Low-Complexity Beam Selection Algorithm for Two-Dimensional Beamspace Massive MIMO

To mitigate the huge investment in radio-frequency devices and associated high power consumption of modern base stations, the beamspace multiple-input multiple-output (B-MIMO) has become a promising solution at an expense of acceptable performance degradation. Facing the increasingly high data throughput demand, two-dimensional (2D) massive B-MIMO systems with beams steering over both elevation and azimuth angles have been developed recently. However, with such a large number of beams available for wireless communication, effective beam selection for multiple users in three-dimensional (3D) space remains technically challenging. In this letter, low-complexity suboptimal beam selection schemes are investigated to overcome this difficulty. By avoiding massive and repetitive matrix-vector product operations in the beam selection procedure through specifically designed linear algebraic calculation strategy, a lower complexity is achieved. Numerical simulations and theoretical analyses demonstrate the proposed methods have around
$K$
to
$K^{2}$
times speedup compared to state-of-the-art beam selection approaches, where
$K$
is the number of users.