A modularized multi-robot SLAM system with elevation mapping and a costmap converter for easy navigation. Different odometry and loop closure algorithms can be easily integrated into the system.
View Demo
·
Report Bug
·
Request Feature
Table of Contents
This is a C++ library with ROS interfaces to manage multi-robot maps. It contains a pluggable front-end FAST-LIO2, pluggable loop closure methods DiSCO / RING. and a global manager that handles submaps, loop candidates and optimization results. The optimizer is mainly based on GTSAM and dist-mapper. The system provides a 3D pointcloud map and an optional 2.5D elevation map output. The output elevation map can be easily converted to a costmap for navigation.
Author: Peter XU (Xuecheng XU)
Affiliation: ZJU-Robotics Lab
Maintainer: Peter XU, [email protected]
Support PCM (Pairwise Consistent Measurement) for robustness.
A chinese version of Tutorial is posted on my blog.
Here, we provide an example to demonstrate the system. Some parameters can be changed to fit your needs.
This software is built on the Robotic Operating System (ROS tested on kinetic and melodic), which needs to be installed first. Additionally, the MR_SLAM depends on following software:
-
Eigen (linear algebra library, tested on 3.2.9 & 3.3.4; elevation_mapping failed on 3.3.9)
-
CUDA (gpu process)
-
Cython (C extensions for Python)
-
GTSAM (pose optimization)
-
Grid Map (grid map library for mobile robots)
sudo apt install ros-$ROS_DISTRO-grid-map*
-
DiSCO (pluggable loop detector)
Follow https://github.com/MaverickPeter/DiSCO-pytorch
-
RING (pluggable loop detector)
Follow the README in LoopDetection/src/RING_ros -
livox_ros_driver (for FAST_LIO2)
Follow https://github.com/Livox-SDK/livox_ros_driver
-
Fast GICP (for ICP refine)
# Fast GICP is already include in the repo. You can use git submodule init --recursive git submodule sync git submodule update # or you can clone the repo and put them in the same place Follow https://github.com/SMRT-AIST/fast_gicp
-
(optional - for elevation_mapping) OctoMap (octomap library for multi-resolution)
sudo apt install ros-$ROS_DISTRO-octomap*
-
(optional - for elevation_mapping) Kindr
Follow https://github.com/anybotics/kindr
- Clone the repo
git clone https://github.com/MaverickPeter/MR_SLAM.git
- Make Mapping
cd Mapping && catkin_make
- Make Localization
cd Localization && catkin_make
- Make LoopDetection
cd LoopDetection && catkin_make # If you encounter the PyInit__tf2 issue, use catkin_make with your python3 environment catkin_make --cmake-args \ -DCMAKE_BUILD_TYPE=Release -DPYTHON_EXECUTABLE=/home/client/miniconda3/envs/py3/bin/python3.8 \ -DPYTHON_INCLUDE_DIR=/home/client/miniconda3/envs/py3/include/python3.8 \ -DPYTHON_LIBRARY=/home/client/miniconda3/envs/py3/lib/libpython3.8.so
-
Get rosbag from Google Drive or Baidu Pan with extract code: gupx
-
Run roscore
-
Run bag
rosbag play 3_dog.bag --clock --pause
-
Run DiSCO / RING
# !!!!! You need to change the Python interpreter to your environment The default is mine: #!/home/client/miniconda3/envs/disco/bin/python3 cd LoopDetection && source devel/setup.bash # DiSCO rosrun disco_ros main.py # RING: If you encounter the PyInit__tf2 issue, use catkin_make with your python3 environment. Check installation section. cd src/RING_ros python main.py # Note that if you add #!/home/client/miniconda3/envs/disco/bin/python3 in the first line of RING_ros/main.py you can also use rosrun RING_ros main.py to start the node.
-
Run global_manager
cd Mapping && source devel/setup.bash roslaunch global_manager global_manager.launch
-
Visualization
rviz -d Visualization/vis.rviz
-
Run roscore
-
Run fast-lio (in 3 terminals)
# Set parameters in Localization/src/FAST_LIO/launch/ and Localization/src/FAST_LIO/config/ !!You need to set the scan_publish_en to true to send submaps cd Localization && source devel/setup.bash roslaunch fast_lio robot_1.launch roslaunch fast_lio robot_2.launch roslaunch fast_lio robot_3.launch
-
Run elevation_mapping (in 3 terminals)
# Set parameters in Mapping/src/elevation_mapping_periodical/elevation_mapping_demos/launch/ and Mapping/src/elevation_mapping_periodical/elevation_mapping_demos/config/ cd Mapping && source devel/setup.bash roslaunch elevation_mapping_demos robot_1.launch roslaunch elevation_mapping_demos robot_2.launch roslaunch elevation_mapping_demos robot_3.launch
-
Run preprocess tools (in 3 terminals)
# If robots don't have cameras, you have to create fake images for elevation_mapping cd Tool/Fake_img python robot_1.py python robot_2.py python robot_3.py # If robots' point cloud are too large for real-time processing, you have to launch filters to accelerate. cd Tool/Filters roslaunch filter_robot_1.launch roslaunch filter_robot_2.launch roslaunch filter_robot_3.launch
-
Run loop detection module
# You need to change the Python interpreter to your environment The default is mine: #!/home/client/miniconda3/envs/disco/bin/python3 cd LoopDetection && source devel/setup.bash # DiSCO rosrun disco_ros main.py # RING cd src/RING_ros python main.py
-
Run global_manager
# Set parameters in Mapping/src/global_manager/launch/ cd Mapping && source devel/setup.bash roslaunch global_manager global_manager.launch
-
Visualization
rviz -d Visualization/vis.rviz
- Add PCM.
- Optimize code.
- Add more pluggable loop closure methods.
- Support more front-end odometry.
See the open issues for a full list of proposed features (and known issues).
If you find this repo useful to your project, please consider to cite it with following bib:
@article{xu2022ring++,
title={RING++: Roto-translation Invariant Gram for Global Localization on a Sparse Scan Map},
author={Xu, Xuecheng and Lu, Sha and Wu, Jun and Lu, Haojian and Zhu, Qiuguo and Liao, Yiyi and Xiong, Rong and Wang, Yue},
journal={arXiv preprint arXiv:2210.05984},
year={2022}
}
Contributions are what make the open source community such an amazing place to learn, inspire, and create. Any contributions you make are greatly appreciated.
If you have a suggestion that would make this better, please fork the repo and create a pull request. You can also simply open an issue with the tag "enhancement". Don't forget to give the project a star! Thanks again!
- Fork the Project
- Create your Feature Branch (
git checkout -b feature/AmazingFeature) - Commit your Changes (
git commit -m 'Add some AmazingFeature') - Push to the Branch (
git push origin feature/AmazingFeature) - Open a Pull Request
Distributed under the MIT License. See LICENSE.txt for more information.
Xuecheng Xu - [email protected]
Project Link: https://github.com/MaverickPeter/MR_SLAM