Human-Centered Artificial Intelligence Application: Recognition of Manual Assembly Movements for Skill-Based Enhancements

Abstract

The vision of Smart Factories, including human-centred intelligent manufacturing systems, have the goal of being highly autonomic systems and aim to assist the human workforce for optimum production stability. It is to be expected that alongside maintenance tasks humans will still perform manual assembly operations for quite a long time. Thus, artificial intelligence methods should be used to assist workers within these tasks. In this paper, we propose the first steps of machine learning method application for recognizing movements of assembly operations for small batch sizes to batch size 1. Therefore, the necessary data collection generated by using activity sensors similar to smart watches or wearables will be described. The generated data will be the bases for the further developments of the AI training. Furthermore, a training concept within a learning factory is briefly introduced.