LIO-Livox: keeping SLAM stable at highway speed and in tunnels

Livox says some users had trouble with high-speed SLAM in busy city traffic, on highways and in dark tunnels, where LiDAR localisation can degrade because the surroundings are plain or open with few features.

• LIO-Livox has two modules. ScanRegistration separates ground, background and foreground points, removes foreground points that are likely moving objects, and picks corner, surface and irregular features.
• PoseEstimation starts the IMU, estimating bias, speed and the direction of gravity, then switches from LiDAR-only to LiDAR-inertial mode. It runs a sliding-window optimisation and builds a global map in a separate thread.
• The source code is published on GitHub as Livox-SDK/LIO-Livox.

• A vehicle drove through a 4 km tunnel at 80 km/h with no loss of quality.
• The system covered more than 10 km of city roads, with high-rise areas, overpasses, highways, tunnels, forest and traffic jams, and Livox reports no failures.
• The point cloud map was registered on Google Maps and reportedly worked well. The article gives no accuracy figures.

One Livox Horizon with a built-in six-axis IMU, mounted on the roof of a demonstration vehicle. The Horizon is an earlier Livox model; for current sensors see our Livox LiDAR range.

• Calibrating several LiDARs to each other, by hand or automatically
• LOAM-Livox: mapping with a Livox Horizon and open-source code