Answers are grounded in this paper and cite their sources.
Context-aware beam management in millimeter-wave (mmWave) wireless communication systems has received increasing attention over the past few years. Machine learning (ML) has played a key role in leveraging different types of context information from the device position and orientation to more ambitious scenarios using RADAR, LIDAR, or camera images. However, most studies in this area consider simplified configurations for user terminals without considering the self-blockage effects owing to the user’s hand and body. This study is a step towards more realistic configurations and scenarios, where methods for location- and orientation-aware beam alignment are evaluated for multi-panel hand-held devices under mild and severe self-blockage conditions. We propose deterministic and probabilistic hand grip schemes that determine the blockage status of device panels. The probabilistic schemes are introduced to account for the inherent randomness of self-blockage owing to variations in the user’s hand grip. Contrary to the blockage models that introduce attenuation in multipath components depending on their angles-of-arrival, we propose two blockage models that introduce blockage losses over all the received paths, which more realistically emulate panels blocked by “hard” hand gripping. Our numerical simulations show that the multi-panel ML-based beam alignment method is capable of leveraging the terminal’s location and orientation information even under severe self-blockage conditions, achieving performance close to genie-aided alignment with just a few beam-pair measurements.
No takes yet. Share an insight, caveat, or question.
Rezaie et al. (2025) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: