Method and apparatus for estimating the position of a moving object using LiDAR scan data

LiDAR-based Mobile Object Localization Using Split Scan Synthesis and Point Cloud Matching

This technology estimates the position of a mobile object by receiving scan data at intervals shorter than the time required for a single LiDAR rotation, synthesizing it with previous scan data to acquire a point cloud, and then identifying the closest point cloud within a point map.

Conventional LiDAR-based localization requires a full sensor rotation to process data, resulting in long position update intervals and often necessitating additional sensors to improve precision.

This technology proposes a pipeline-based approach that synthesizes and matches scan data received at short intervals, shortening the localization cycle without the need for extra sensors. It can be applied to autonomous vehicles and indoor logistics robots, providing an economical solution that enhances both position update speed and precision using only existing sensors.

‍

‍

Key Features:

• Receiving scan data at a second duration that is shorter than the first duration required for a single LiDAR rotation
• Synthesizing the received scan data with scan data from a previous time point stored in units of the second duration
• Acquiring a first point cloud of the mobile object's surroundings based on the synthesized scan data
• Estimating the position based on the point cloud in the point map that is closest to the first point cloud

‍

This invention was developed with the support of the Ministry of Science and ICT's research on fault-tolerant real-time virtualization technology for high-reliability autonomous driving systems.