Precueing Sequential Rotation Tasks in Augmented Reality

Augmented reality has been used to improve sequential-task performance by cueing information about a current task step and precueing information about future steps. Existing work has shown the benefits of precueing movement (translation) information. However, rotation is also a major component in many real-life tasks, such as turning knobs to adjust parameters on a console. We developed an AR testbed to investigate whether and how much precued rotation information can improve user performance. We consider two unimanual tasks: one requires a user to make sequential rotations of a single object, and the other requires the user to move their hand between multiple objects to rotate them in sequence. We conducted a user study to explore these two tasks using circular arrows to communicate rotation. In the single-object task, we examined the impact of number of precues and visualization style on user performance. Results show that precues improved performance and that arrows with highlighted heads and tails, with each destination aligned with the next origin, yielded the shortest completion time on average. In the multiple-object task, we explored whether rotation precues can be helpful in conjunction with movement precues. Here, using a rotation cue without rotation precues in conjunction with a movement cue and movement precues performed the best, implying that rotation precues were not helpful when movement was also required.