The beam alignment problem in mmWave wireless networks

Abstract

Millimeter-wave (mmWave) cellular systems are a promising option for a very high data rate connectivity in near future because of the large bandwidths available at this range of frequencies. Communication at mm Waves, however, is quite challenging due to the large path-loss exponent compared with the traditional sub-6GHz counterpart. Directional beamforming with large antenna gains both at the user and at the base-station is necessary to cope with this problem. Designing such a highly directional beamforming requires a reliable estimate of the channel state such as angle-of-arrivals (AoAs) of dominant scatterers in the channel connecting a user and a base-station. This is, however, quite challenging to obtain due to the very low signal-to-noise ratio available before suitable beamforming. In this paper, we study interactive beam-alignment algorithms by which the user and the base-station collaborate to identify the AoA of at least one strong scatterer over a mmWave channel. We specify a theoretical framework for assessing the performance of different beam-alignment algorithms. We also propose a parametric family of Markovian strategies and analyze its achievable performance theoretically and empirically via numerical simulations. Our theoretical analysis indicates the key channel parameters governing the performance of such a Markovian beam-alignment algorithm.