The use of different modes of transportation to traverse urban environments is a common technique for humans. For autonomous robots, this poses a variety of challenges, as their navigation architectures are typically designed to handle only one specific navigation task. To overcome this problem in urban navigation scenarios, we present a skill-based navigation architecture that is capable of switching between different navigation modes and utilizes different modes of transportation, including public transportation. By representing the urban environment as a skill graph, it is possible to draw from a set of skills for different navigation challenges, including sidewalk navigation, road crossing, and riding public transportation. Furthermore, this representation can be automatically extracted using geospatial data. With the proposed navigation architecture, the robot can efficiently traverse urban environments by autonomously walking to a public transportation station, waiting for the public transportation vehicle, riding it to the desired stop, and then walking to the desired destination. The proposed architecture is validated in a navigation scenario through the city of Graz, Austria on the legged robot platform ANYmal. By extracting a skill graph from OpenStreetMap data, the robot can autonomously plan and execute a route to its desired destination using public transportation.
Enabling Multimodal Mobility for Urban Robot Navigation Using A Skill-Based Navigation Architecture
2024-09-24
2873744 byte
Conference paper
Electronic Resource
English
The Robotability Score: Enabling Harmonious Robot Navigation on Urban Streets
ArXiv | 2025
|IEEE | 2015
|Vertiport navigation requirements and multisensor architecture considerations for urban air mobility
Springer Verlag | 2024
|