CoLRIO

A ROS2 package of CoLRIO: LiDAR-Ranging-Inertial Centralized State Estimation for Robotic Swarms.

9.21.6._202422975822.mp4

Prerequisites

• Ubuntu ROS2 Foxy (Robot Operating System 2 on Ubuntu 20.04)
• CMake (Compilation Configuration Tool)
• PCL (Default Point Cloud Library on Ubuntu work normally)
• Eigen (Default Eigen library on Ubuntu work normally)
• GTSAM 4.2a8 (Georgia Tech Smoothing and Mapping library)

Compilation

Build CoLRIO:

mkdir -p ~/cslam_ws/src
cd ~/cslam_ws/src
git clone https://github.com/PengYu-Team/Co-LRIO.git
cd ../
colcon build --symlink-install

Run with Dataset

• [our dataset] TBD.

• S3E dataset. The datasets are configured to run with default parameter.

ros2 launch co_lrio run.launch.py
ros2 bag play *your-bag-path*

Citation

This work is published in IEEE ICRA 2024 conference, and please cite related papers:

@misc{zhong2024colrio,
      title={CoLRIO: LiDAR-Ranging-Inertial Centralized State Estimation for Robotic Swarms}, 
      author={Shipeng Zhong and Hongbo Chen and Yuhua Qi and Dapeng Feng and Zhiqiang Chen and Jin Wu and Weisong Wen and Ming Liu},
      year={2024},
      eprint={2402.11790},
      archivePrefix={arXiv},
      primaryClass={cs.RO}
}

@article{feng2022s3e,
  title={S3e: A large-scale multimodal dataset for collaborative slam},
  author={Feng, Dapeng and Qi, Yuhua and Zhong, Shipeng and Chen, Zhiqiang and Jiao, Yudu and Chen, Qiming and Jiang, Tao and Chen, Hongbo},
  journal={arXiv preprint arXiv:2210.13723},
  year={2022}
}

Acknowledgement

• We combined the front end of CoLRIO and the DLO to achieve the 5th position in the ICCV 2023 LiDAR-Inertial SLAM Challenge.

The Leaderboard is shown as follow:

And the hardware and results are shown as follow:

• CoLRIO depends on FAST-GICP (Kenji Koide, Masashi Yokozuka, Shuji Oishi, and Atsuhiko Banno, "Voxelized GICP for fast and accurate 3D point cloud registration".).

• CoLRIO depends on GncOptimizer (Yang, Antonante, Tzoumas, Carlone, "Graduated Non-Convexity for Robust Spatial Perception: From Non-Minimal Solvers to Global Outlier Rejection").