Projekt 9: Service robots with trustworthy situational decisions

Intelligent service robots must perform tasks in open environments and in workspaces shared with humans. For successful task execution, service robots need to adequately and autonomously decide on their respective course of action based on the situation and the task. In this project, a closed chain from data-based training for recognition and classification of complex situations (pictures of various situations or use of further data channels, statistical data) to discrete clusters with adequate influence on action alternatives (semantic coupling into discrete actions) shall be realized. A prerequisite for the use and acceptance of learning methods, however, is the explainability and comprehensibility of the coupling of situational images and action decisions. Emerging benefits would be explainable reactions and differentiated actions of service robots in everyday situations for robust and safe task execution.

First supervisor:

Prof. Dr. Christian Schlegel, Technische Hochschule Ulm

 

Tandem partner:

Prof. Dr. Birte Glimm, Institut für Künstliche Intelligenz, Universität Ulm

 

Consulting experts:

Prof. Dr. Hans Kestler, Institut für Medizinische Systembilologie, Universität Ulm

Prof. Dr. Michael Munz, Technische Hochschule Ulm