TL;DR
This paper introduces TeamPortal, a novel VR collaboration system that allows users to share and manipulate each other's field of view (FOV) through a parallel view window. Studies show that TeamPortal significantly improves task efficiency and social presence in complex collaborative tasks by enabling direct object manipulation across shared perspectives, offering a more intuitive way to transfer items than traditional methods.
What Problem Does This Paper Address?
Collaborative Virtual Environments (CVEs) are increasingly used for remote work and socialization, but transferring virtual objects between users remains a challenge. Existing methods often struggle with identifying target objects, especially when they are outside a user's direct line of sight or occluded. This can lead to difficulties in coordination, increased movement, and reduced collaboration efficiency. The paper addresses this by proposing a system that allows users to not only see their partner's perspective but also interact with objects within that shared view, aiming to streamline object transfer and enhance overall collaboration.
Methods
The core innovation is TeamPortal, a system that presents a semi-transparent, real-time view of a collaborator's FOV within a user's own VR environment. This "parallel view" window can be moved and made opaque for clearer interaction. Crucially, users can use their VR controllers to "shuttle" into this view window and directly grab and manipulate objects, either transferring them to their own FOV or placing them into the collaborator's view. A "locking mechanism" ensures only one user can manipulate an object at a time to prevent conflicts.
The authors also developed and evaluated three variants:
- TeamPortal+: Introduced optimizations for perspective synchronization, triggering updates only when head movement exceeds certain thresholds to reduce motion sickness and network load.
- SnapTeamPortal+: Allowed users to "pause" the view window, freezing the collaborator's perspective for easier interaction.
- DropTeamPortal+: Similar to SnapTeamPortal+, but the paused view appeared in a separate, temporary window.
Two user studies were conducted to compare these techniques against a baseline (no shared view) and a simpler shared view (ShaView).
Datasets
Not discussed in the paper. The studies used custom-designed collaborative tasks within a virtual environment.
Key Insights and Ideas
- Manipulation is Key: Simply seeing a partner's view (ShaView) is less effective than being able to actively manipulate objects within that view (TeamPortal), especially in complex tasks.
- Task Complexity Matters: TeamPortal's benefits are most pronounced in complex tasks where object identification and transfer are more challenging. For simple tasks, the overhead of using the parallel view can sometimes be detrimental.
- Pausing for Precision: The ability to pause the shared view (SnapTeamPortal+, DropTeamPortal+) can significantly improve task efficiency and user preference by allowing for more deliberate interactions without the pressure of real-time synchronization.
- Trade-offs in Design: Different users have different priorities. Some prefer the efficiency of pausing (SnapTeamPortal+), while others value continuous co-presence (TeamPortal+). This suggests a need for flexible interface designs.
- Reducing Movement: TeamPortal significantly reduces user movement and teleportation in complex environments, which can also help mitigate VR-induced disorientation.
Key Results
User Study 1 (TeamPortal vs. Baseline vs. ShaView): * Complex Tasks: TeamPortal significantly improved task completion rates and accuracy compared to Baseline and ShaView. It also reduced movement and teleportation. * Simple Tasks: No significant differences were found in task completion or accuracy. TeamPortal even showed a slight tendency to increase completion time, possibly due to added cognitive load. * Subjective Measures: TeamPortal led to significantly better overall user experience, pragmatic quality, and hedonic quality. It also significantly enhanced co-presence and perceived behavioral interdependence. * Simulator Sickness: All techniques resulted in similar, minimal simulator sickness symptoms.
User Study 2 (TeamPortal+ vs. SnapTeamPortal+ vs. DropTeamPortal+): * Complex Tasks: SnapTeamPortal+ and DropTeamPortal+ increased task completion rates compared to TeamPortal+. * User Preference: SnapTeamPortal+ and DropTeamPortal+ were generally preferred over TeamPortal+. * Co-presence: SnapTeamPortal+ slightly reduced co-presence compared to the other variants. * User Feedback: Some users found the two-window design of DropTeamPortal+ cognitively demanding, while others appreciated the ability to pause. SnapTeamPortal+ was polarizing, with some finding it efficient and others missing the continuous co-presence.
Limitations and Future Work
- VR Experience: Participants in both studies had limited VR experience, which might have influenced their perception of usability and preferences. Future work could explore how varying levels of VR experience affect system adoption.
- Comparison with Visual Cues: The study did not directly compare TeamPortal with systems that use visual cues like annotations or pointing, which are effective when users' FOVs overlap. Future research could integrate these.
- 2D Rendering: The current implementation uses 2D rendering for view windows, potentially losing depth information. Exploring real-time stereoscopic rendering for parallel views is a future direction, though it may increase bandwidth demands.
- Multiplayer Scenarios: Extending TeamPortal to support more than two collaborators is a significant future direction, raising questions about how to organize and display multiple parallel views without overwhelming users.
- Design Implications: The authors propose three design implications: offering multiple view window modes, adapting to task complexity, and considering the size of the collaborative environment.
Summary updated