Industrial Augmented Reality – Feasibility of AI-based Validation of AR Instructed Component Recovery for Washing Machines in Circular Economy

Using Artificial Intelligence (AI) to validate that tasks are performed correctly when instructed through Augmented Reality (AR) is a rather unexplored area. Focusing on the removal of components during the disassembly of washing machines in the context of circular economy, we develop an AR system, to detect correctly executed tasks using AI. We present preliminary results of our novel approach using the YOLOv8 Nano object model detecting eight components. Already with very limited training data we could achieve detection precisions of over 90% for seven of eight components proofing the general feasibility of our work.