Design and Evaluation of AR-Assisted End-User Robot Path Planning Strategies

BSTRACT Nowadays, robots play an increasingly important role but still usually have to be programmed by highly skilled professionals. Therefore, end-user solutions supporting users in solving (simple) robot programming tasks without expert knowledge are a promising research field. A possibility for these solutions is the inclusion of Augmented Reality (AR) to enable users to work in the robot space, reducing the amount of mentally taxing coordinate space conversions. Approaches for this mainly rely on the waypoint-based robot path programming strategy. To explore an alternative solution, we propose an AR-assisted approach with different possible path-planning strategies, such as drawing paths or selecting single waypoints directly in the real world. This enables end-users without a programming background to program paths for a wheeled mobile robot. They can also see and edit their programmed paths in a Blockly-like representation. Furthermore, we offer AR in-place program simulation and direct building of finished programs to the real robot. We evaluated our approach regarding usability compared to existing non-AR end-user software as well as comparing the two different path-planning strategies to each other. The evaluation showed that our approach is more usable and faster than the conventional method, while the differences between the path-planning methods are more nuanced and show qualities in both versions.