In-situ or side-by-side? A user study on augmented reality maintenance instructions in blind areas

In recent years, Augmented Reality (AR) has been effectively proposed as a tool to support workers in manual procedural tasks in industry, such as assembly and maintenance. It is very common for workers to deal with complete equipment in maintenance. Then, instructions may refer to components located in blind areas, i.e., visually occluded by part of the equipment. Displaying in-situ AR instructions in such blind areas requires handling occlusions; otherwise, side-by-side instructions can be exploited. In the literature, it is still unclear which solution better leverage workers' performance in maintenance tasks accomplished in blind areas. Thus, we designed three AR presentation modes to convey maintenance instructions in a real machine: 3D in-situ rendered with X-ray technique; 3D side-by-side on a CAD model replicating the blind area; 2D side-by-side on a virtual mirror. We conducted a user study comparing these three presentation modes with a 2D drawing extracted from the original maintenance documentation of the machine. The performance of 42 participants was evaluated in terms of completion time, accuracy, and cognitive load. The results revealed that both in-situ and 3D side-by-side presentation modes perform better than 2D drawing. Specifically, the in-situ presentation mode outperforms the 3D side-by-side mode in terms of completion time. The side-by-side virtual mirror does not improve performance with respect to 2D drawing, then it needs to be redesigned for effective use in AR maintenance interfaces.