Full solutions for automatic transportation vehicles in unstructured environmentsare restricted to military applications, although possible mission scenarios exist forcivil and space applications. In this thesis, a user-friendly, human-robot interfaceis implemented for an all-terrain electric vehicle. The interface allows the vehicleto follow a moving user autonomously by combining 3D LIDAR measurements withintensity images. The method detects clusters within the LIDAR scan that matchthe human appearance and resolves potential ambiguities based on image trackingresults. The interface can be controlled using different input devices. A detailedevaluation compares the input devices, analyzes the implemented algorithms, andvalidates the overall system behavior outdoor under real-world conditions. Thedeveloped system is able to follow its user as long as he is completely visible. ; Validerat; 20140822 (global_studentproject_submitter)
Development and Evaluation of a Human-Robot Interface for an Autonomous All-Terrain Transportation Vehicle in order to Localize, Track and Follow a Human.
2014-01-01
Local fe9f2c57-4708-4ad7-b83c-9e389d95ebb3
Hochschulschrift
Elektronische Ressource
Englisch
DDC: | 629 |