Adaptive Beam Search for Initial Beam Alignment in Millimetre-Wave Communications

In this correspondence, we propose an adaptive beam search algorithm to solve the initial beam alignment problem in millimetre-wave communications, where there is neither prior knowledge on the instantaneous channel fading status nor the fading statistics. The proposed algorithm works by performing iterative beam scanning. In each iteration, a candidate beam set that likely contains the strongest beam are identified by examining the posterior probability that one beam is stronger than another under an improper prior consideration. We then propose a new scanning strategy that selects one beam for scanning from the candidate beam set in each iteration and develop a beam decision method to identify the final beamforming direction. Numerical results show that the proposed algorithm can achieve better beam alignment accuracy with lower time complexities, compared to existing adaptive search algorithms.