Lidar-level localization with radar? The CFEAR approach to accurate, fast and robust large-scale radar odometry in diverse environments
🏆 Our article has won the "Best Paper Award in Computer & Robot Vision!"
News (January 2023) CFEAR is integrated in TBV Radar SLAM
News (April 2023): Article is now published in T-RO, code is released
- A video demo of our latest results is shown here.
CFEAR-3 journal: "Lidar-level localization with radar? The CFEAR approach to accurate, fast and robust large-scale radar odometry in diverse environments"
- Paper: T-RO or arXiv
- A demo is found here. The video intends to visually demonstrate the experiments carried out in the paper
- We release most of our content, including
- Our diverse radar datasets
- Our full evaluation which includes source code for paper figures
- Source code
citation
@ARTICLE{9969174,
author={Adolfsson, Daniel and Magnusson, Martin and Alhashimi, Anas and Lilienthal, Achim J. and Andreasson, Henrik},
journal={IEEE Transactions on Robotics},
title={Lidar-Level Localization With Radar? The CFEAR Approach to Accurate, Fast, and Robust Large-Scale Radar Odometry in Diverse Environments},
year={2023},
volume={39},
number={2},
pages={1476-1495},
doi={10.1109/TRO.2022.3221302}}
CFEAR-2 Conference article: CFEAR Radar odometry - Conservative Filtering for Efficient and Accurate Radar Odometry
Presented at IROS 2021
- Published conference article and preprint
- IROS 2021 presentation
citation
@INPROCEEDINGS{9636253, author={Adolfsson, Daniel and Magnusson, Martin and Alhashimi, Anas and Lilienthal, Achim J. and Andreasson, Henrik},
booktitle={2021 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)},
title={CFEAR Radarodometry - Conservative Filtering for Efficient and Accurate Radar Odometry},
year={2021},
volume={},
number={},
pages={5462-5469},
doi={10.1109/IROS51168.2021.9636253}}
- The initial work on CFEAR was presented at Radar Perception for All-Weather Autonomy, a Half-Day Workshop at 2021 IEEE International Conference on Robotics and Automation (ICRA)
- Workshop preprint
- Workshop presentation
citation
@article{DBLP:journals/corr/abs-2109-09994,
author = {Daniel Adolfsson and Martin Magnusson and Anas W. Alhashimi and Achim J. Lilienthal and Henrik Andreasson},
title = {Oriented surface points for efficient and accurate radar odometry},
journal = {CoRR}, volume = {abs/2109.09994}, year = {2021}, url = {https://arxiv.org/abs/2109.09994}, eprinttype = {arXiv}, eprint = {2109.09994},
timestamp = {Mon, 27 Sep 2021 15:21:05 +0200},
biburl = {https://dblp.org/rec/journals/corr/abs-2109-09994.bib},
bibsource = {dblp computer science bibliography, https://dblp.org}
}
TBV Radar SLAM - trust but verify loop candidates
Large-scale introspective SLAM using Navtech radar.
Bibtex
@ARTICLE{10103570,
author={Adolfsson, Daniel and Karlsson, Mattias and Kubelka, Vladimír and Magnusson, Martin and Andreasson, Henrik},
journal={IEEE Robotics and Automation Letters},
title={TBV Radar SLAM – Trust but Verify Loop Candidates},
year={2023},
volume={8},
number={6},
pages={3613-3620},
doi={10.1109/LRA.2023.3268040}}
Learns detection of small localization errors using Navtech radar.
Bibtex
@article{DBLP:journals/corr/abs-2109-09994,
author = {Daniel Adolfsson and Martin Magnusson and Anas W. Alhashimi and Achim J. Lilienthal and Henrik Andreasson},
title = {Oriented surface points for efficient and accurate radar odometry},
journal = {CoRR}, volume = {abs/2109.09994}, year = {2021}, url = {https://arxiv.org/abs/2109.09994}, eprinttype = {arXiv}, eprint = {2109.09994},
timestamp = {Mon, 27 Sep 2021 15:21:05 +0200},
biburl = {https://dblp.org/rec/journals/corr/abs-2109-09994.bib},
bibsource = {dblp computer science bibliography, https://dblp.org}
}
- If you have any questions, feel free to contact me: Daniel Adolfsson (dan11003) [email protected]