Approaches for Cognitive Assistance in Industry 4.0 Assembly

The high degree of customization and small lot sizes necessitate a considerable amount of manual assembly work in industry 4.0 scenarios. Cognitive assistance systems could enable workers with limited experience to fulfill these roles. Approaches for designing such systems are illustrated by ADAMAAS-I (“Adaptive and Mobile Action Assistance for Industry”), an advanced cognitive assistance system using smart glasses in the context of industry 4.0 production scenarios. The aim of ADAMAAS-I is to move from stationary systems to a mobile and adaptive action support and monitoring system that reacts dynamically to human actions and prevent mistakes by giving appropriate assistance. For this purpose, ADAMAAS-I uses 3D in-situ instructions via augmented reality and state-of-the-art machine learning techniques for action recognition. Furthermore, ADAMAAS-I incorporates a systematic and largely automatized method from cognitive science for analyzing users’ action representations in long-term memory. This information about individual memory structures is used to anticipate human errors and to improve the accuracy of action recognition. By these means, ADAMAAS-I provides individualized, user- and situation-specific assistance, e.g. for manual assembly tasks in an industrial context.