Improving Human-Robot Interaction Through Decision Support and Workplace-Based Learning: Prototype of a Worker Assistance System for Adaptive Task Sharing Between Robots and Humans

Abstract

Collaborative robots (cobots) enable flexible cooperation between humans and robots. There are different approaches for the division of labor, on the one hand the static approaches, where tasks are assigned once to either the human or the robot according to better suitability or technical feasibility, and on the other hand the Adaptive Task Sharing approach (ATS). In ATS, so-called shareable tasks are assigned by the worker to the human or robot based on situation dependent requirements, for example, worker preferences, quality, lot size, etc. However, this ATS approach also brings new issues. The dynamic assignment at the worker level creates a multitude of possible process variants, making it difficult for the worker to find the best one for the current situation. Existing ATS implementations in human-robot interaction, however, are currently only concerned with the feasibility of the concept, neglecting the topic of decision support and learning. Based on the Design Science Research methodology, an existing web-based ATS prototype of the TU Wien Pilot Factory 4.0 was significantly extended with decision support and learning capabilities to support and relief workers in collaborative assembly tasks. The implementation of the enhanced prototype and the results of the evaluation focusing on the usability and the user experience are shown in this paper. Therefore, this paper contributes to the research on the development and implementation of decision support and learning capabilities with a focus on cobot technology in learning factories.