Jeon, Jinhwan (Korea Advanced Institute of Science and Technology(KAIST)), Lim, Chaeho (Korea Advanced Institute of Science and Technology(KAIST)), Hwang, Yoonjin (Korea Advanced Institute of Science and Technology(KAIST), Hanko), Choi, Seibum (KAIST)

Detection and Classification of Driving Maneuver Events for Automatic Lane Map Generation

Scheduled for presentation during the Regular Session "Data Management and Geographic Information Systems" (FrBT2), Friday, September 27, 2024, 13:50−14:10, Salon 5

2024 IEEE 27th International Conference on Intelligent Transportation Systems (ITSC), September 24- 27, 2024, Edmonton, Canada

This information is tentative and subject to change. Compiled on December 26, 2024

Abstract

Collecting data through Mobile Mapping System(MMS) and manually processing it to generate lane-level maps is a laborious and resource-intensive process. To address this challenge, recent advancements have introduced methods for automatically generating maps from collected data. However, most studies have addressed simple driving scenarios or manually excluded specific driving maneuver trajectories during the map-generation process. Additionally, they have focused on creating custom maps rather than adhering to established formats like Lanelet2. This paper presents a novel algorithm for automatic map generation in the Lanelet2 format. Our approach involves a two-stage cascade classifier designed to detect and classify vehicle driving maneuvers, facilitating the segmentation of vehicle trajectories into multiple lanelets. To assess the effectiveness of our approach, we conducted evaluations using sensor data gathered from real vehicle test drives in Sejong City, South Korea.