Location-Free Beam Prediction in mmWave Systems

Tushara Ponnada*, Hanan Al-Tous, Olav Tirkkonen

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference article in proceedingsScientificpeer-review

2 Citations (Scopus)
160 Downloads (Pure)


Channel charting is a method for creating radiomaps of a cell that capture the neighborhood relationships between User Equipments (UEs) in the cell based on machine learning techniques. In this paper, we leverage channel charting
for predicting the best Base Station (BS) beam to serve a given UE in a massive-MIMO 5G network. Because of the autonomous beamforming at the UE in 5G networks, the BS cannot determine the best beam for transmission to a UE by measuring the UE transmissions in all the BS beams. To address this issue, we
propose a framework to predict the best BS beam for a mobile UE in the next transmission instant by utilizing the channel charts of the cell that the UE is currently in. We evaluate the prediction accuracy of the framework using simulated channels from QuaDRiGa channel generator. We compare the performance of channel chart and physical location based predictors. While
the prediction accuracy attained using channel charting is less than that of the prediction using physical locations, there remain several ways to improve the performance.
Original languageEnglish
Title of host publicationProceedings of IEEE 93rd Vehicular Technology Conference, VTC 2021
Number of pages6
ISBN (Electronic)978-1-7281-8964-2
Publication statusPublished - 15 Jun 2021
MoE publication typeA4 Conference publication
EventIEEE Vehicular Technology Conference - Helsinki, Finland
Duration: 25 Apr 202128 Apr 2021
Conference number: 93

Publication series

NameIEEE Vehicular Technology Conference
ISSN (Print)1090-3038
ISSN (Electronic)2577-2465


ConferenceIEEE Vehicular Technology Conference
Abbreviated titleVTC-Spring


  • mmWave
  • CSI features
  • 5G TDD system
  • channel charting
  • BS beam prediction


Dive into the research topics of 'Location-Free Beam Prediction in mmWave Systems'. Together they form a unique fingerprint.

Cite this