Abstract With recent advancements in technology, augmented reality (AR) systems are now capable of providing practical shared experiences offering new ways to work, socialize, and play together. While previous research has explored systems for asynchronous collaboration, studies focusing on real-time co-located synchronous collaboration where two or more users perform the same task together in tandem remain limited. Additionally when users are working in shared collaborative environments together, the level of symmetry to their movements (i.e., symmetric motions in their paired coordination) and the requirements for asymmetry (i.e., individualize motions within the collaboration) is presently underexplored. Therefore while highly immersive, realistic collaborative XR systems can be created, it is unclear how the symmetry provided by the system, notably in the collaborative interaction paradigms, should be designed to optimize user experience and task performance. To address this, we present a formal user study of two common AR interaction paradigms which offer varying levels of symmetry in a paired task. We explore Grasp (symmetric) and AirTap (asymmetric) within a bi-manual, co-located manipulation task where users (n = 32) in pairs (n = 16) complete a collaborative AR docking task. To illustrate the level of successful collaboration and user experience, we report on multiple metrics, notably task completion time, spatial accuracy, errors (drops in interaction), team workload and usability. Furthermore to explore user opinion and negate the over reliance on post study user questionnaires, we apply a sentiment analysis process directly to the communication between user pairs during the collaborative task. This supports a richer exploration of user positive or negative sentiment between the Grasp (symmetric) and AirTap (asymmetric) conditions. Results demonstrate that the level of symmetry offered by the interaction paradigm (Grasp and AirTap) presents significant differences in collaborative performance, team workload and usability while maintaining comparable task spatial accuracy. Task performance is found to be lower with grasping interactions which required greater symmetry between user motions and actions and resulted in longer completion times and higher collaborative effort. Sentiment of the paired user communication further highlighted a more reliable collaboration in the asymmetric (AirTap) condition. These findings align to prior work assessing asymmetry in interaction and therefore support the suggestion that symmetric interaction methods (Grasp), while intuitive, may hinder coordination and increase cognitive demands paired collaboration. We recommend that AR researchers and designers of collaborative virtual environments consider introducing levels of controlled asymmetry in interaction paradigms to better support user performance.
Khattri et al. (Mon,) studied this question.