Developing a Taxonomy of Spatial Information for Augmented Reality Training: Design Strategies to Enhance Procedural Activity Performance

Augmented reality (AR) training has the potential to enhance learning by overlaying virtual content onto physical spaces and representing abstract concepts through multisensory cues. However, AR designers face challenges in determining the optimal amount and types of information to enhance learning without causing cognitive overload. This paper proposes a spatial information taxonomy developed through activity component analysis, which decomposes tasks into fundamental behavioral processes and provides a systematic approach to identifying essential information types and their representations in AR training applications. By analyzing prior AR training scenarios using video coding methods, the authors identified key information types—such as instruction, interaction guidance, and feedback—and corresponding representation forms, including Panel, Glyph, Ghost, Trajectory, and others. The taxonomy offers a structured framework to guide designers in integrating pedagogically effective spatial information. Future work will validate the taxonomy and examine conditions under which specific information types and representations best optimize learning outcomes.