EUVIMEDEUVIMED entdecken
Crossref

Radar-Aided Beam Prediction and Tracking: Will It Work in the Real World?

Umut Demirhan, Ahmed Alkhateeb

IEEE Transactions on Communications · 2026 · 74 · 4336-4352

DOI: 10.1109/tcomm.2026.3655783 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Transactions on Communications, 2026. Verfasst von Umut Demirhan und 1 weiteren Autor:innen. Die Arbeit wurde bisher 4-mal zitiert (Crossref) und nennt 48 Literaturangaben. Gefördert von: National Science Foundation.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Millimeter wave (mmWave) and terahertz (THz) communication systems require precise beam alignment, but the extensive beam training process creates significant overhead, especially challenging for latency-sensitive and highly mobile applications like vehicular and augmented reality. To address this critical issue, this paper presents an in-depth machine learning study on radar-aided mmWave beam prediction and tracking. For the beam prediction task, we develop and evaluate various machine learning-based solutions, utilizing preprocessing techniques on the radar signals to extract relevant features. For beam tracking, we propose a modular approach integrating classical radar tracking with an extended Kalman filter and machine learning classification, alongside various end-to-end deep learning architectures leveraging CNNs, ResNet, LSTMs, and transformers to directly learn the mapping from radar data to optimal beams. The proposed methods are assessed using the large-scale real-world DeepSense 6G dataset, which provides co-existing mmWave beam training and radar measurements in a realistic vehicular communication scenario. Our extensive comparative analysis across various settings demonstrates the effectiveness of the proposed approaches, achieving high beam prediction accuracy while significantly reducing beam training overhead, and yields critical insights into the performance characteristics of different machine learning models for radar-aided beam prediction and tracking, providing valuable guidance for the development of advanced sensing-assisted beam management techniques in mobile networks.

Förderung

  • National Science Foundation

Zitieren

APA

Demirhan, U., & Alkhateeb, A. (2026). Radar-Aided Beam Prediction and Tracking: Will It Work in the Real World?. IEEE Transactions on Communications, 74, 4336-4352. https://doi.org/10.1109/tcomm.2026.3655783

Vancouver

Demirhan U, Alkhateeb A. Radar-Aided Beam Prediction and Tracking: Will It Work in the Real World?. IEEE Transactions on Communications. 2026;74:4336-4352. doi:10.1109/tcomm.2026.3655783

BibTeX
@article{Demirhan2026,
  author = {Demirhan, Umut and Alkhateeb, Ahmed},
  title = {Radar-Aided Beam Prediction and Tracking: Will It Work in the Real World?},
  journal = {IEEE Transactions on Communications},
  year = {2026},
  volume = {74},
  pages = {4336--4352},
  doi = {10.1109/tcomm.2026.3655783}
}
Kostenlos starten →