Exploiting OpenStreetMap-Data for Outdoor Robotic Applications

Abstract

The application of (semi-)autonomous outdoor robots for delivery or transport purposes promises flexible, comfortable and cost-efficient solutions. Beside autonomous driving on an automotive level, smaller robotic systems are going to conquer footpaths and cycle tracks. The individual success of those applications depends on the geographical nature of the operational area (global profile of heights, availability of suitable tracks with a specific width without stairs or street bollards) and legal restrictions for autonomous robots (permissions for street types, speed limits or required street crossing mechanisms). Both aspects together determine the chances for a successful operating robotic scenario and have to be evaluated related to the consequences - travel time, energy consumption, potential risks. This paper describes a tool for outdoor applications analyzing public data from the Open Street Map project. Based on a critical assessment of relevant data related to completeness and quality it presents a chain of data aggregation, contextual filtering and graph generation according to specific legal restrictions. For evaluation purposes, the paper defines a set of metrics providing comparable key indicators and applies the concept to an exemplary delivery robot scenario with different sets of driving restrictions.