Map building using helmet-mounted LiDAR for micro-mobility.

Area of Science:

• Robotics and Autonomous Systems
• Geospatial Data Science

Background:

• Accurate 3D mapping is crucial for micro-mobility navigation.
• Motion and vibration from micro-mobility vehicles cause significant distortion in LiDAR data.

Purpose of the Study:

• To develop a robust point-cloud mapping method for micro-mobility.
• To correct motion-induced distortions in LiDAR measurements for improved mapping accuracy.

Main Methods:

• Utilized a helmet-mounted Light Detection and Ranging (LiDAR) sensor.
• Employed Normal Distributions Transform-based Simultaneous Localization and Mapping (NDT SLAM) and an Inertial Measurement Unit (IMU) for pose estimation.
• Applied a Kalman filter-based algorithm for motion distortion correction, assuming near-constant velocity motion.
• Mapped corrected LiDAR data onto an elevation map and extracted stationary objects using the occupancy grid method.

Main Results:

• Successfully corrected LiDAR measurement distortions caused by micro-mobility motion.
• Generated accurate point-cloud maps of the environment by utilizing stationary object measurements.
• Demonstrated the effectiveness of the proposed method in a campus road environment.

Conclusions:

• The proposed method effectively builds point-cloud maps for micro-mobility by correcting LiDAR motion distortion.
• This approach enhances the reliability and accuracy of mapping in dynamic environments.
• The integration of NDT SLAM, IMU, and Kalman filtering provides a robust solution for micro-mobility mapping.