Effect of Gaze Visualization on Task Efficiency and User Behavior for Guidance Scenarios in Co-Located AR Collaboration

Augmented Reality (AR)-enabled collaboration enhances contextual guidance in shared environments, where unambiguous cues are essential, for swift and accurate task completion. Gaze-based cues offer clear, non-verbal indication of intent, reducing interaction ambiguity. While prior research has studied the effectiveness of gaze-based visualization in scenarios where collaborators share the same level of knowledge, less attention has been shown to guidance scenarios, where one user provides task-related information to another. Existing studies often focus on task outcomes (e.g., completion time, error rate) without analyzing gaze data to gain further insight into user behavior and attention. This study leverages gaze-based cues as an active means of communication in guidance scenarios, assessing their impact on both task performance and user behavior through gaze data analysis. The developed system incorporates gaze-based visualizations, including dynamic representations of gaze behavior, to deliver task-critical guidance. These are compared to verbal instructions using quantitative metrics for task effectiveness and gaze analysis, alongside subjective feedback, to determine whether they reduce ambiguity and enhance collaboration. Our study showed that gaze-based visualizations significantly reduced task related errors, with contextual triggered cues offering the best balance between speed and correctness. Gaze analysis revealed that gaze-based visualizations effectively directed participants’ attention to task-relevant areas, reducing unnecessary visual exploration. Participants reported lower perceived workload in gaze-based conditions, further supporting their effectiveness in enhancing collaboration and reducing ambiguity.