Home // International Journal On Advances in Systems and Measurements, volume 13, numbers 3 and 4, 2020 // View article

Point Cloud Mapping and Merging in GNSS-Denied and Dynamic Environments Using Onboard Scanning LiDAR

Authors:
Seiya Tanaka
Chisato Koshiro
Misato Yamaji
Masafumi Hashimoto
Kazuhiko Takahashi

Keywords: LiDAR; point cloud map; mapping and merging; NDT-Graph SLAM

Abstract:
This paper presents a 3D point cloud mapping and merging in Global Navigation Satellite Systems (GNSS)-denied and dynamic environments using only a scanning Light Detection And Ranging (LiDAR) mounted on a vehicle. Distortion in scan data from the LiDAR is corrected by estimating the vehicle’s pose (3D positions and attitude angles) in a period shorter than the LiDAR scan period using Normal Distributions Transform (NDT) scan matching and Extended Kalman Filter (EKF). The corrected scan data are mapped onto an elevation map. Static and moving scan data, which originate from static and moving objects, respectively, in the environments are classified using the occupancy grid method. Only the static scan data are utilized to generate several submaps in different small areas using NDT-based Simultaneous Localization And Mapping (NDT SLAM) and Graph SLAM. These submaps are merged using Graph SLAM. Experimental results obtained in outdoor residential and urban road environments show the LiDAR-based mapping and merging via EKF and NDT-Graph SLAM provide accurate maps in GNSS-denied and dynamic environments.

Pages: 275 to 288

Copyright: Copyright (c) to authors, 2020. Used with permission.

Publication date: December 30, 2020

Published in: journal

ISSN: 1942-261x