FastLivo2

The FAST-LIVO2 module integrates LiDAR–Inertial–Visual Odometry for UAV navigation
and mapping. In the MAVISS stack, FAST-LIVO2 is configured for the Livox Mid-360,
providing robust state estimation and mapping in GPS-denied or cluttered environments.
This module plays a similar role as RSFastLivo (for the RoboSense AC1),
but optimized for Livox sensors.

Repository

Features

  • LiDAR–Inertial–Visual fusion for odometry.

  • Compatible with the Livox Mid-360 sensor.

  • Real-time mapping suitable for UAV navigation and SLAM research.

  • Robust in low-texture and GPS-denied environments.

  • ROS/ROS2 integration for visualization in RViz.

Installation & Usage

Clone, build, and launch FAST-LIVO2 in your ROS 2 workspace:

# Create a workspace
mkdir -p ~/ros2_ws/src && cd ~/ros2_ws/src

# Clone the FAST-LIVO2 repository
git clone https://github.com/hku-mars/FAST-LIVO2.git

# Build
cd ~/ros2_ws
colcon build --packages-select fast_livo2
source install/setup.bash

# Launch FAST-LIVO2 with example launch file
ros2 launch fast_livo2 run.launch.py

Integration Notes

  • Requires Livox Mid-360 + IMU + camera for full VIO pipeline.

  • Ensure proper calibration between LiDAR, camera, and IMU frames.

  • Provides odometry output that can be fused with PX4 via MAVROS or uXRCE-DDS.

  • Recommended for indoor and outdoor UAV missions where GPS may be unreliable.