🔥 This is a slim version of Super Odometry, containing the LiDAR Odometry component and IMU Odometry component. The LiDAR odometry only provides pose constraints to IMU odometry modules to estimate the bias of IMU. In return, the IMU Odometry module offers pose predictions to the LiDAR Odometry module, serving as an initial guess for ICP optimization.
🔥 The system has been widely tested on above platforms equipped with Livox, Velodyne and Ouster LiDAR.
- Multi-LiDAR Support
- Compatible with Livox, Velodyne, and Ouster sensors
- LiDAR-inertial Fusion
- Support LiDAR-inertial Fusion
- Dual-Mode Operation
- Supports both localization and mapping modes
- Alignment Risk Prediction
- Provides alignment risk prediction for ICP algorithms
- Degeneracy Awareness
- Robust detection of environmental degeneracy
- ROS 2.0 Integration
- Built on ROS 2 Humble for modern robotics development
🔥 6 DOF degeneracy uncertainty detection. We support visualization in both RVIZ and Rerun.
Highly recommend to check our docker files to run our code with step 4 and step 5.
- ROS2 Humble
- PCL
- Eigen
- Sophus
- GTSAM (4.0.2 or 4.1)
- Ceres Solver (2.1.0)
git clone http://github.com/strasdat/Sophus.git
cd Sophus && git checkout 97e7161
mkdir build && cd build
cmake .. -DBUILD_TESTS=OFF
make -j8 && sudo make installgit clone https://github.com/borglab/gtsam.git
cd gtsam && git checkout 4abef92
mkdir build && cd build
cmake \
-DGTSAM_USE_SYSTEM_EIGEN=ON \
-DGTSAM_BUILD_WITH_MARCH_NATIVE=OFF \
..
make -j6 && sudo make installgit clone https://github.com/ceres-solver/ceres-solver.git
cd ceres-solver
git checkout f68321e7de8929fbcdb95dd42877531e64f72f66
mkdir build
cd build
cmake ..
make -j8 # Use number of cores you have, e.g., -j8 for 8 cores
sudo make installpip install rerun-sdkcd ros2_humble_docker
docker build -t superodom-ros2:latest .First create your own local ROS2 workspace and clone SuperOdom:
mkdir -p ~/ros2_ws/src
cd ~/ros2_ws/src
git clone https://github.com/superxslam/SuperOdomClone respective repos and ensure they follow this exact structure under ros2_ws/src:
ros2_ws/src
├── SuperOdom
├── livox_ros_driver2
└── rviz_2d_overlay_plugins
You can clone livox_ros_driver2 and rviz_2d_overlay_plugins using the following link:
Important: Maintain this exact structure within
ros_ws/src
# Allow Docker GUI access
xhost +local:dockerGo to ros2_humble_docker/container_run.sh and make sure you change exact directory path for PROJECT_DIR and DATASET_DIR
PROJECT_DIR="/path/to/your/superodom"
DATASET_DIR="/path/to/your/dataset"Important:
PROJECT_DIRshould be the exact directory toros2_ws/src
Then launch docker container using the following:
# Grant access
cd ros2_humble_docker
sudo chmod -R 777 container_run.sh
# Start container
./container_run.sh superodom-ros2 superodom-ros2:latest
# Source ROS2
source /opt/ros/humble/setup.bashImportant: To access container, you can open a new bash window and run
docker exec --privileged -it superodom-ros2 /bin/bash
Build the workspace within container
cd ~/ros2_ws/src/livox_ros_driver2
./build.sh humble
cd ~/ros2_ws
colcon buildImportant: make sure you first build
livox_ros_driver2
To launch SuperOdometry, we provide demo datasets for Livox-mid360, VLP-16 and OS1-128 sensor Download Link
For more challange dataset, feel free to download from our website slam_mode and localization_mode. You might want to convert ROS1 bag into ROS2 format using this link.
For user-defined topic names, modify super_odometry/config/$(YOUR_LiDAR_SENSOR).yaml:
imu_topic: "/your/imu/topic"
laser_topic: "/your/laser/topic"For user-defined laser-imu extrinsics, modify super_odometry/config/$(YOUR_LiDAR_SENSOR)/$(YOUR_LiDAR_SENSOR)_calibration.yaml:
#Rotation from laser frame to imu frame, imu^R_laser
extrinsicRotation_imu_laser: !!opencv-matrix
rows: 3
cols: 3
dt: d
data: [1., 0., 0.,
0., 1., 0.,
0., 0., 1.]
#Translation from laser frame to imu frame, imu^T_laser
extrinsicTranslation_imu_laser: !!opencv-matrix
rows: 3
cols: 1
dt: d
data: [-0.011, -0.02329, 0.04412]Run SuperOdometry using the following command:
source install/setup.bash
ros2 launch super_odometry livox_mid360.launch.py
ros2 launch super_odometry os1_128.launch.py
ros2 launch super_odometry vlp_16.launch.pyPlay your ROS2 dataset:
# launch this in a new bash window
docker exec --privileged -it superodom-ros2 /bin/bash
source install/setup.bash
cd ~/data
ros2 bag play $(YOUR_ROS2_DATASET)Visualize in RVIZ2:
# launch this in a new bash window
docker exec --privileged -it superodom-ros2 /bin/bash
source install/setup.bash
cd ~/ros2_ws/src/SuperOdom/super_odometry
rviz2 -d ros2.rviz(⭐ Alternative) Visualize in Rerun:
# launch this in a new bash window
docker exec --privileged -it superodom-ros2 /bin/bash
source install/setup.bash
cd ~/ros2_ws/src/SuperOdom/script/visualizers
python3 rerun_visualizer.py
# Open a new bash window on your local device
rerunWe also provide tmux script for easy launch with dataset (this script only works after you build the workspace in docker):
cd script
tmuxp load run.yamlcave_website1.mp4
🔥 The localization mode allows you to localize your robot by providing an initial pose and ground truth map.
Update your super_odometry/config/$(YOUR_LiDAR_SENSOR).yaml configuration file with:
localization_mode: true # If true, localization mode is enabled; otherwise, SLAM mode is used
read_pose_file: false # Set to true to read initial pose from a txt file
init_x: 0.0 # Initial X position for localization
init_y: 0.0 # Initial Y position for localization
init_z: 0.0 # Initial Z position for localization
init_roll: 0.0 # Initial roll angle
init_pitch: 0.0 # Initial pitch angle
init_yaw: 0.0 # Initial yaw angleAdd ground truth map map in launch file
parameters=[LaunchConfiguration("config_file"),
{ "calibration_file": LaunchConfiguration("calibration_file"),
"map_dir": os.path.join(home_directory, "/path/to/your/pcd"),
}]To quickly launch our localization module, feel free to try out this demo dataset using default initial pose configuration.
@inproceedings{zhao2021super,
title={Super odometry: IMU-centric LiDAR-visual-inertial estimator for challenging environments},
author={Zhao, Shibo and Zhang, Hengrui and Wang, Peng and Nogueira, Lucas and Scherer, Sebastian},
booktitle={2021 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)},
pages={8729--8736},
year={2021},
organization={IEEE}
}
@inproceedings{zhao2025superloc,
title={SuperLoc: The Key to Robust LiDAR-Inertial Localization Lies in Predicting Alignment Risks},
author={Zhao, Shibo and Zhu, Honghao and Gao, Yuanjun and Kim, Beomsoo and Qiu, Yuheng and Johnson, Aaron M. and Scherer, Sebastian},
booktitle={2025 IEEE International Conference on Robotics and Automation (ICRA)},
year={2025},
url={https://arxiv.org/abs/2412.02901}
}🔵 Colorized Point Cloud Visualization — Video Demo
🟢 Visual Odometry Module — Initial Release Lightweight and robust visual odometry module integrated into SuperOdometry.
This package is released under the GPLv3 license. For commercial use, please contact shiboz@andrew.cmu.edu and Prof. Sebastian Scherer.
Special thanks to Professor Ji Zhang, Professor Michael Kaess, Parv Maheshwari, Yuanjun Gao, Yaoyu Hu for their valuable advice. Thanks to Omar Alama for providing Rerun support. We also acknowledge these foundational works: