Fast Beam Search for Massive MIMO Based on Mainlobe Overlapping State of Training Beam

Abstract

To reduce beam search complexity without decreasing feedback efficiency for millimeter-wave communication in the LoS environment, this paper proposes a fast beam search scheme based on joint judgment, which requires a masterly design of the training beams and let their mainlobe overlap according to certain rules. As a consequence, its state utilization efficiency has been improved to 100% while keeping feedback efficiency is still 100%. In this paper, through theoretical analysis, we find that to encode the index of every subinterval $A_{q}$ using the Gray mapping can decrease the possibility and the impact of misestimating, compared with using a binary index, which is adopted. The transceiver emits the training beam in turn and then jointly determines optimal communication beam pair according to the relationship between the receiving power and the threshold. The simulation results show that our proposed scheme is more efficient and its search complexity has been further decreased while its FE remains 100%. Especially, this method has more obvious advantages in multi-user simultaneously beam search scenarios.