POINT-LIO-SLAM

This repository provides ready-to-use ROS 2 Humble configuration and launch files for composing FAST-LIO-SAM-QN and FAST-LIO-Localization-QN (original repositories) along with their respective ROS 2 wrappers (FAST-LIO-SAM-QN and FAST-LIO-Localization-QN). These configurations are tailored for UniLidar L1 indoor mapping and localization. No new algorithmic code is introduced; instead, this repository packages configuration files, launch scripts, and RViz presets to enable seamless integration and one-command bring-up of the upstream projects.

Dependencies

• C++ >= 17
• OpenMP >= 4.5
• CMake >= 3.10.0
• Eigen >= 3.2
• Boost >= 1.54
• ROS 2 Humble (Ubuntu 22.04 recommended)
• GTSAM
• TEASER++
• TBB (used by Quatro for speed)
• Livox-SDK2 (for UniLidar L1)

Install GTSAM (4.1.1)

wget -O gtsam.zip https://github.com/borglab/gtsam/archive/refs/tags/4.1.1.zip
unzip gtsam.zip
cd gtsam-4.1.1/
mkdir build && cd build
cmake -DGTSAM_BUILD_WITH_MARCH_NATIVE=OFF -DGTSAM_USE_SYSTEM_EIGEN=ON ..
sudo make install -j16

Install TEASER++

git clone https://github.com/MIT-SPARK/TEASER-plusplus.git
cd TEASER-plusplus && mkdir build && cd build
cmake .. -DENABLE_DIAGNOSTIC_PRINT=OFF
sudo make install -j16
sudo ldconfig

Install Livox-SDK2

git clone https://github.com/Livox-SDK/Livox-SDK2.git
cd ./Livox-SDK2/
mkdir build
cd build
cmake .. && make -j
sudo make install

Install TBB

sudo apt install -y libtbb-dev

Workspace layout

Create a ROS 2 workspace and clone the required repositories under src/:

mkdir -p ~/my_ws/src && cd ~/my_ws/src
# This repo
git clone https://github.com/mvu20002/POINT-LIO-SLAM.git
# Upstream components
git clone https://github.com/mvu20002/Localization-QN.git
git clone https://github.com/mvu20002/SAM-QN.git
git clone https://github.com/illusionaryshelter/Quatro.git
git clone https://github.com/illusionaryshelter/nano_gicp.git
git clone https://github.com/mvu20002/livox_ros_driver2_humble.git
git clone https://github.com/dfloreaa/point_lio_ros2.git

Build

cd ~/my_ws
source /opt/ros/humble/setup.bash
colcon build --symlink-install
source install/setup.bash

Run

NOTE: The configuration parameters are set for UniLidar L1 on a mobile robot platform horizontally mounted on the robot. Data used for mapping and localization couldn't be shared due to privacy concerns. However, the configuration files are provided for you to adapt to your own data.

Prepare Unitree Bag

Official UniLidar L1 bag files can be downloaded from here. Since the bag file in ROS 1 format is not compatible with ROS 2, you need to convert it to ROS 2 format. You can use this tool for conversion.

Mapping

To run the mapping process, run the following command:

ros2 launch point-lio-slam mapping.launch.py
# another terminal
ros2 bag play <path_to_ros2_bag>

When you end the mapping process, you can save the map by pressing Ctrl+C in the terminal where the mapping is running. The map will be saved in the ~/my_ws/install/sam-qn/share/sam_qn/ directory with the name result.bag. (Check the original repository for more details on the mapping process.)

Localization

Once you have the map, you can use it for localization. Change the path to map in config file localization_indoor.yaml to the path of your saved map. Then, run the following command:

ros2 launch point-lio-slam localization.launch.py