A Two-Stage Fast Beam Search for Multi-User Millimeter-Wave Massive MIMO Systems

Abstract

Millimeter-wave systems deploy spatial beam search for initial access. A narrow beam search results in significant discovery delays whereas wide beams cause high user misdetection. In this work, a novel two-stage fast beam search algorithm is proposed that utilizes a beamforming architecture with an appropriate antenna spacing for subarray design. The first stage employs analog beamforming to identify the coarse beam space. The second stage based on hybrid beamforming makes use of pattern multiplication for beam space refinement. Simulation results demonstrate that the proposed scheme achieves improved performance in comparison to existing techniques in terms of significantly reduced beam search latency with a marginal increase in user misdetection.