Design Aspects for Efficient Beam Training in 6G Millimeter-Wave Networks with Mobile Users

With 6G millimeter-wave (mmWave) networks aiming to make their services available to mobile users, beam management procedures need to be reevaluated. Many studies have focused on link maintenance for mobile users already connected to the network, however, little is known about the implications of mobile users intending to access the network. Beam training design trade-offs in terms of data rate, delay, and coverage area have been observed in studies with stationary users. This work builds on this state-of-the-art by investigating the differences between stationary and mobile initial access along these metrics. We find that the exhaustive search algorithm's optimal performance degrades rapidly for mobile users by selecting stale beam pairs, for example, resulting in tens of decibel power loss and, particularly, failure in more than 80 % of the initial access at 45 km/h owing to the high search duration. This problem can be shifted to higher velocities by using broader beams for lower delay, but at the cost of decreased coverage. Instead, our work highlights that beam training for mobile users should terminate immediately once a beam pair of a desirable quality is found to avoid staleness. By lowering the target threshold of our proposed scheme, the delay is minimized at the cost of selecting less-robust links. However, we demonstrate that there is a sweet spot that minimizes delay while maximizing the link strength, and thereby also enabling users at typical urban velocities to access the network at a negligible failure risk by minimizing loss.