SPARK-FAST-LIO

LiDAR mapping = SPARK-FAST-LIO + KISS-Matcher-SAM

---

📦 How to Install

Put the code in your workspace/src folder

cd ${YOUR_COLCON_WORKSPACE}/src
git clone https://github.com/MIT-SPARK/spark-fast-lio.git
colcon build --packages-up-to spark_fast_lio

🚀 How to Run

We provide two out-of-the-box ROS2 examples using pre-processed ROS2 bag data (because the original data are only available in ROS1). All pre-processed ROS2 bag files can be found here.

🇺🇸 LIO on the MIT campus

1. Download 10_14_acl_jackal and 10_14_hathor (from the Kimer-Multi dataset)

2. Run spark_fast_lio using the following command:

ros2 launch spark_fast_lio mapping_mit_campus.launch.yaml scene_id:=acl_jackal

3. In another terminal, run ROS2 bag file as follows:

ros2 bag play 10_14_acl_jackal

🏟️ LIO on the Colosseum

1. Download colosse_train0 (from the VBR dataset)

2. Run spark_fast_lio using the following command:

ros2 launch spark_fast_lio mapping_vbr_colosseo.launch.yaml

3. In another terminal, run ROS2 bag file as follows:

ros2 bag play colosseo_train0

🔧 How to run spark-fast-lio2 using your own ROS2 bag?

1. Copy config/velodyne_mit.yaml or config/ouster_vbr.yaml to config/${YOUR_CONFIG}.yaml, and set the appropriate values for:
   • lidar_type, scan_line, timestamp_unit, and filter_size_map depending on your sensor type
   • extrinsic_T and extrinsic_R (it's LiDAR w.r.t. IMU, i.e., extrinsic * cloud w.r.t. LiDAR -> cloud w.r.t. IMU)
2. Configure your launch file and remap the lidar and imu topic names to match your setup.
   • Also set an appropriate rviz setup
3. Run your launch file, for example: ros2 launch spark_fast_lio ${YOUR_LAUNCH}.launch.yaml

How to run with KISS-Matcher-SAM?

Please carefully read README.md of KISS-Matcher-SAM before running the command.

1. To install kiss_matcher_ros in your colcon workspace, run:

cd ${YOUR_ROS2_WORKSPACE}/src
git clone https://github.com/MIT-SPARK/KISS-Matcher.git
cd ..
colcon build --packages-select kiss_matcher_ros

2. Then, run the command below:

ros2 launch kiss_matcher_ros run_kiss_matcher_sam.launch.yaml

3. By default, this setup is compatible with the two examples above (i.e., the topics are already remapped to support them). However, if you want to run it on your own dataset, make sure to set the /cloud and /odom topics appropriately using:

ros2 launch kiss_matcher_ros run_kiss_matcher_sam.launch.yaml \
  odom_topic:=<YOUR_TOPIC> scan_topic:=<YOUR_TOPIC>

---

What's New? Key Features and Updates

1. Complete code refactoring for better structure and readability

2. Visualization frame can now be configured via launch file

3. Added support for gravity alignment

---

Acknowledgement

Thanks to HKU MaRS Lab guys. The original code is from FAST-LIO2, which can be found here.