Hash-based Fast Beam Alignment for 6G Sub-Terahertz MIMO

Abstract

The 6th generation mobile networks require large bandwidth on sub-Terahertz bands to provide more than 100 Gbps throughput. To provide enough beamforming gain to combat the path loss on high-frequency bands, large-scale antenna arrays should be used, which generate narrow and directive beams. It may require hundreds or thousands of beams to cover a cell and introduce challenges for the beam alignment during the initial access stage. As a result, both the overhead and latency of beam sweeping may become unacceptable. We propose the Hash-based beam alignment method to address this problem, including a Hash-based beam transmission scheme for the synchronization signal block (SSB) transmissions and a corresponding UE beam pair selection method. It utilizes multiple transceivers to transmit multiple beams simultaneously, where the beam order follows a pattern generated by a Hash function. Considering UE may only receive discrete SSBs randomly, we design a weighted voting algorithm for UE to fast select the optimal beam pair for the following data transmissions. Simulations demonstrate that the proposed methods can reduce about 50% of the beam alignment latency or find a better beam pair with about 3 dB higher beamforming gain given a limited searching time.