G. C. F. Lee, E. Kurniawan, and Y. Zheng, "Odometry-Aided mmWave Communications using Overparameterized Beamforming Optimization,& IEEE Vehicular Technology Conference 2024 Spring (IEEE VTC 2024 Spring), In-Proceedings, Singapore, June 2024.
Abstract:
In vehicular-to-infrastructure (V2I) applications, as we adopt higher frequencies and antenna array processing for faster data rates, beamforming is a critical enabling technology to preserve high-quality communication links between mobile users and the base station. Traditional beamforming requires overhead in establishing the directional link, compounded by short coherence time of mobile users, which necessitates frequent re-calibration. Recent works consider more efficient strategies by leveraging context information, with position-aware approaches showing particular promise. Nonetheless, uncertainties in even the most advanced localization and odometry methods can lead to compounded errors in position estimation for downstream tasks such as beam alignment and tracking. Our work addresses these challenges by factoring in the uncertainties inherent in practical odometry systems. We propose a beamforming optimization that accounts for this uncertainty, complemented by overparameterization strategies in gradient-based optimization to the non-linear problem of interest. Our simulation results demonstrate the feasibility of maintaining robust communication links in representative scenarios by leveraging localization and odometry information, despite the challenges posed by a mobile platform with imprecise position estimates.
License type:
Publisher Copyright
Funding Info:
This research / project is supported by the National Research Foundation / Infocomm Media Development - Future Communications Research & Development Programme
Grant Reference no. : FCP-NTU-RG-2022-021