A comparison of surface and motion user-defined gestures for mobile augmented reality.

Abstract

Augmented Reality (AR) technology permits interaction between the virtual and physical worlds. Recent advancements in mobile devices allow for a better mobile AR experience, and in turn, improving user adoption rate and increasing the number of mobile AR applications across a wide range of disciplines. Nevertheless, the majority of mobile AR applications, that we have surveyed, adopted surface gestures as the default interaction method for the AR experience and have not utilised three-dimensional (3D) spatial interaction, as supported by AR interfaces. This research investigates two types of gestures for interacting in mobile AR applications, surface gestures, which have been deployed by mainstream applications, and motion gestures, that take advantages of 3D movement of the handheld device. Our goal is to find out if there exists a gesture-based interaction suitable for handheld devices, that can utilise the 3D interaction of mobile AR applications.

We conducted two user studies, an elicitation study and a validation study. In the elicitation study, we elicited two sets of gestures, surface and motion, for mobile AR applications. We recruited twenty-one participants to perform twelve common mobile AR tasks, which yielded a total of five-hundred and four gestures. We classified and illustrated the two sets of gestures, and compared them in terms of goodness, ease of use, and engagement. The elicitation process yielded two separate sets of user-defined gestures; legacy surface gestures, which were familiar and easy to use by the participants, and motion gestures, which found to be more engaging. From the design patterns of the motion gestures, we proposed a novel interaction technique for mobile AR called TMR (Touch-Move-Release). To validate our elicited gestures in an actual application, we conducted a second study. We have developed a mobile AR game similar to Pokémon GO and implemented the selected gestures from the elicitation study. The study was conducted with ten participants, and we found that the motion gesture could provide more engagement and better game experience.

Nevertheless, surface gestures were more accurate and easier to use. We discussed the implications of our findings and gave our design recommendations for designers on the usage of the elicited gestures. Our research can be further explored in the future. It can be used as a "prequel" to the design of better gesture-based interaction technique for different tasks in various mobile AR applications.