µ-Lanelet graph for precise global path planning of robo-taxi

초록

Unlike self-driving buses, which have fixed stops, robo-taxis require precise global path planning to reach a target location quickly and accurately. Precisely generating a global path using a Lanelet graph and a starting/ending point graph, proposed in previous studies, are inefficient because they do not consider a position within the lane. Herein, we propose a µ-Lanelet graph for precise global path planning of robo-taxis and a method for converting a high definition roadmap provided by the National Geographic Information Institute into a Lanelet map. In a µ-Lanelet graph, each lane of a Lanelet map is divided into several sections and connected to each other. It is possible to generate an optimal drivable route directly considering multi-lane changes and construction sites only by searching the graph. The validity of the proposed method was verified through a test using a real vehicle in K-City, Hwaseong, Republic of Korea. © 2021, Institute of Control, Robotics and Systems. All rights reserved.

키워드