This paper considers the problem of path-planning for long-term operation of autonomous mobile robots in changing environments. As a deterministic behaviour of the robot is required in many applications, a path is planned using a predefined route map. Due to the open-ended operation, the route representation may become out of date after some time. To solve this problem, an approach inspired by human memory information processing is proposed which maintains the map up-to-date based on the environmental observations. Additionally, statistical data like travel times and route blockages are gathered and integrated in the path planner daytime-dependant to improve the overall routing performance. Following the description of the adaptive path planner, experimental results demonstrate the functionality of the presented approach. Section II discusses previous work on navigation approaches in changing environments. Section III introduces the route map, routing algorithm and navigation methods this work bases on. Section IV presents the adaptive path planner for long-term navigation followed by the contribution to long-term routing in Section V. Experimental results are presented in Section VI. Finally, the conclusion and discussion for future work is presented in Section VII.
Adaptive path planning for long-term navigation of autonomous mobile robots
2009
6 Seiten, 4 Bilder, 3 Tabellen, 20 Quellen
Aufsatz (Konferenz)
Englisch