Skillab - A Multimodal Augmented Reality Environment for Learning Manual Tasks

Abstract

The paper investigates the usage of AR-based systems in teaching manual skills. We propose Skillab, a novel AR-based scaffolding system. It aids in the learning of manual work and functions as a multimodal immersive tool for feedback, including muscle actuation. As our initial investigation, we made a floor lamination tutorial. We evaluated our system’s performance and user experience in comparison to traditional paper instructions. With 20 participants, we conducted a between-group user study and obtained both qualitative and quantitative data. In terms of task performance, learnability, and overall user experience, Skillab significantly outperformed conventional paper instructions. In contrast to learning from paper instructions, Skillab training demonstrated a significant improvement in the systematic rating on the quality of the performed task. We believe that by demonstrating the potential of immersive multi-modal feedback technology for skill-building, researchers would be motivated to explore this area further.