LAUSR

Combined with massive multiple-input multiple-output technology, beamforming can effectively compensate the high path-loss of millimeter-wave signals and hence improve the received power at user equipment (UE). In order to establish the beam alignment between base station (BS) and UE, beam training (BT) is required to select appropriate beamforming vectors at both BS and UE sides. In general, a good trade-off between beamforming gain and BT overhead is critical to enhance the system performance. In this letter, a statistics-assisted beam training (SA-BT) scheme is proposed to reduce the BT overhead. Specifically, BS scans the candidate beams according to the descending order of the utilization frequency of each beam. Meanwhile, the BT procedure is finished when the beamforming gain at UE exceeds the pre-defined threshold. The theoretical analyses and the numerical simulations prove that the proposed scheme can reduce the BT overhead and hence improve the spectral efficiency significantly compared with the existing methodologies.