Beam Training and Beam Tracking in Millimeter Wave Systems

Availability of large bandwidth in millimeter wave (mmWave) frequencies can lead to multi-Gbps data rates in next generation wireless networks. High path loss and severe shadowing are among the major concerns to achieve such data rates in mmWave bands. Various beamforming (BF) techniques can be applied to cope with these effects by forming directional transmission patterns called narrow beams. Furthermore, several experimental studies confirm that mmWave channel incorporates a few spatial clusters. Consequently, it is essential to devise beam training algorithms for aligning the corresponding narrow beams with the direction of channel, i.e. the angle of arrival (AoA) associated with spatial clusters of the channel.

In order to enable reliable data communication through narrow beams, the base station (BS) and user are required to allocate a fraction of the time to estimate the AoA through beam training, and the remaining time to data communication. On the one hand, spending more time on the estimation of AoA leads to narrower beams which increases the BF gain during data communication, but on the other hand less time remains for data transmission. In this project, one of the objectives is to study this trade off and design optimized procedures for joint beam training and data transmission in mmWave systems so as to maximize the system utility (e.g. average data rate). The optimization includes finding optimal beam training protocols and the optimal duration of beam training and data communication phases.

While the problem is easier to tackle in static scenarios where the user location is fixed and the propagation channel is line of sight (LoS), it is more complicated to find the optimal beam training and data communication strategies where the user is mobile and/or we have a multipath propagation environment. Another objective of this project is to devise high performance beam tracking algorithms using which the BS and user can track the long-term variations of the channel caused by user mobility. Furthermore, we aim to investigate the problem of joint beam training/tracking and data communication in multi-user scenarios.