Skip to content

Latest commit

 

History

History
23 lines (18 loc) · 2.3 KB

README.md

File metadata and controls

23 lines (18 loc) · 2.3 KB

A Novel Look at LIDAR-aided Data-driven mmWave Beam Selection

This repo contains the code for the paper "A Novel Look at LIDAR-aided Data-driven mmWave Beam Selection" avilable at: https://arxiv.org/abs/2104.14579

Authors

Dataset

In this work we use Raymobtime dataset, which is a collection of realistic ray-tracing data obtained by simulating traffic in environment highly inspired by real world data. It utilizes SUMO for mobility simulations, Insite for ray-tracing, Cadmapper and Open Street Map for importing realistic outdoor scenarios. The dataset is divided into smaller sets, with different frequencies considered, various number of receivers, and environments. We trained our models on the s008 dataset (we combined both training and validation subsets) and validated on s009.

Questions?

If you have any further questions related to this repo, feel free to contact us at zecchin@eurecom.fr or raise an Issue within this repo. We will do our best to reply as soon as possible.

References

  1. Matteo Zecchin, Mahdi Boloursaz Mashhadi, Mikolaj Jankowski, Deniz Gunduz, Marios Kountouris, David Gesbert, “A Novel Look at LIDAR-aided Data-driven mmWave Beam Selection”, arXiv:2104.14579v1 [cs.IT] 29 Apr 2021, link: https://arxiv.org/abs/2104.14579
  2. ITU AI/ML in 5G Challenge website: https://www.itu.int/en/ITU-T/AI/challenge/2020/Pages/default.aspx
  3. Raymobtime dataset: https://www.lasse.ufpa.br/raymobtime/
  4. A. Klautau, P. Batista, N. González-Prelcic, Y. Wang and R. W. Heath Jr., “5G MIMO Data for Machine Learning: Application to Beam-Selection using Deep Learning” in 2018 Information Theory and Applications Workshop (ITA).