Development of an Augmented Reality Interface for Intuitive Robot Programming

Abstract

This thesis is dedicated to the research of a feedback mechanism in the context of learning from demonstration, which is intended to improve the interaction between humans and robots. In this context, an augmented reality interface for the HoloLens 2 in combination with the SARA robot was developed and subsequently evaluated in a user study with 20 participants. The aim was to investigate the influence of feedback during the learning from demonstration process on the intuitive interaction with the robot, the efficiency of the demonstrations, the reduction of the knowledge gap between man and machine and the quality of the demonstrations. The results show that augmented reality interfaces have the potential to improve the intuitive handling of the robot. Furthermore, they have the potential to increase the user's understanding of the robot's learning process. However, no specific quality improvements were found in the results of algorithm learning from demonstrations. The discussion of the results emphasizes the importance of user preferences and requirements in the development of augmented reality interfaces for learning from demonstration systems. Future research will focus on investigating more complex tasks, alternative output devices and different interaction methods for non-expert users. These findings will help to better understand how augmented reality technologies can improve the approach of learning by demonstration and make it more intuitive for non-experts.