On Multilevel Road Mapping for Autonomous Vehicles: A Study to Generate Accurate 2.5D LIDAR Maps Using Graph SLAM in Challenging Environments

点击翻译按钮获取中文摘要

I. A unique tactic to automatically map multilevel environments and localize autonomous vehicles accordingly in each layer has been reported in a Journal paper (Sensor: Remote Sensing).II.Many collected mapping data were analyzed to determine the challenging environments to generate accurate maps using GNSS/INS-RTK systems and investigate the reasons. This study was very important to be a reference to investigate and collect mapping data in the future. High buildings, dense, trees, long tunnels, multilayer junctions, bridges and underpasses were found to be the most critical road structures to deflect and obstruct the satellite signals received by GNSS/INS-RTK systems. Accordingly, the importance of generating accurate maps in multilevel road structures has been emphasized and the study and relevant analysis have been reported in a Journal paper (Sensor: Remote Sensing). III. A unique strategy to fix the global position inconsistency in the XY plane between road layers has been highlighted at The 25th IEEE International Conference on Intelligent Transportation Systems (IEEE ITSC 2022). IV. The capability to precisely combine/merge maps in the XY plane using the proposed Graph SLAM framework has been analyzed and integrated into the mapping module. The relevant results have been published in The 2022 IEEE/ASME International Conference on Advanced Intelligent Mechatronics (AIM).