Adaptive Mensch-Roboter-Kollaboration für die ergonomische Gestaltung von Montagearbeit

• Adaptive human-robot collaboration for the ergonomic design of assembly work

Petruck, Henning; Nitsch, Verena (Thesis advisor); Meisen, Tobias (Thesis advisor)

Düren : Shaker Verlag (2021)
Book, Dissertation / PhD Thesis

In: Industrial engineering and ergonomics 41
Page(s)/Article-Nr.: xiv, 199 Seiten : Illustrationen, Diagramme

Dissertation, RWTH Aachen University, 2021

Abstract

Increasing individualization and shorter product lifecycles hinder the economic realization of full automation in production. Therefore, flexible production systems are increasingly being used, which often include collaborative robots that work together with humans in production. This dissertation focusses the use case of human-robot collaboration in assembly, where the robot assists the human by picking up components. For this use case an ergonomic design of the interaction by user adaption was developed. The term user adaptation refers to the temporal adjustment of the handover times to the working speed of the human being. Based on the assumption that an early handover of the component has a negative effect on trust in the robot and leads to increased mental strain, while a late handover can lead to unwanted delays in production, the adaptive concept was developed in such a way that the robot hands over the component as precisely as possible at the time it is needed. This objective was implemented by designing an adaptive concept tailored to the application and by developing a prediction algorithm for predicting the execution times of manual assembly steps. In this thesis the development of different prediction algorithms was described and their suitability for the application case was evaluated in a simulation study. Based on the results of the simulation study, the most suitable of the algorithms was selected for use in further studies. In addition to the use of diverse the prediction algorithms, it was also discussed under which conditions data from a production network can be used for prediction using statistical methods. Finally, the designed adaptation concept was implemented and evaluated in the context of a subject study. In the subject study, the participants were given screws which they then had to screw into a workpiece. The robot adapted to the working speed in one condition, and as a comparison there was a second test condition in which the robot did not adapt to the working speed. The study aimed at investigating the effects of adaptation on the subjective perception of the test subjects and on objective measures of the human-robot team performance. Neither for the subjective perception of the subjects, nor for the total execution time and number of errors, effects of the adaptation could be identified. However, there were significant effects for the waiting times, especially the waiting time of the robot could be significantly reduced for slowly working subjects.

Institutions

• Internet of Production [080067]
• Chair and Institute of Industrial Engineering and Ergonomics [417110]