Selection in Stride: Comparing Button- and Head-Based Augmented Reality Interaction During Locomotion

Military users of augmented reality (AR) head-mounted displays must interact with their heads-up displays (HUDs) effectively and while on the move. Yet, there is a paucity of human-computer interaction (HCI) studies investigating AR multimodal interfaces (MMIs) and interaction methods during locomotion. We conducted a mixed methods study comparing stationary and ambulatory button- and head-based AR interaction methods. Utilizing a within-participants design, Soldier participants completed a simple task sequence in an AR HUD while walking on an omnidirectional treadmill and standing still using both a chest-mounted controller alone (C) and a head-gaze cursor with button input for selection (C + HG). Quantitative task performance analysis revealed faster time-on-task for the C + HG method when stationary. However, when walking, the C method generally surpassed the C + HG method. Careful analysis of selection and head-gaze hovering inputs reflected participants’ difficulty in stabilizing their head while walking which led to inaccuracies in menu icon selection and necessitated additional selection input. Moreover, several participants reported difficulty with stabilizing their head-gaze as well as greater preference for and better success using the C method to perform the task sequence while walking. Taken together, these findings support the idea that while head-gaze is a promising AR interaction method in relatively stationary contexts, the fact that it requires good head stability for reliable interaction negatively impacts task performance and user experience during locomotion. This study brings attention to the challenges of MMIs in ambulatory AR usage contexts and the need for more research in this area.