Partially Blended Realities: Aligning Dissimilar Spaces for Distributed Mixed Reality Meetings

Title of the Paper

Partially Blended Realities: Aligning Dissimilar Spaces for Distributed Mixed Reality Meetings

Paper Information

• Research Area: Mixed Reality Systems and Remote Collaboration
• Keywords: Mixed Reality, Augmented and Virtual Reality, Remote Collaboration, Mixed Reality Spaces, Proximal Transitions

Research Background and Problem Statement

• Identified Challenges:

  • Mixed reality technology has the potential to align distributed physical spaces virtually, creating a shared interactive space for remote collaboration. However, aligning these spaces becomes challenging when users are in dissimilar physical environments.
  • Traditional approaches often require highly consistent physical environments, which is incompatible with the dynamic nature of collaborative work.
  • In heterogeneous spaces, users cannot maintain continuous spatial alignment when transitioning between different surfaces, potentially disrupting the collaborative experience.

• Significance:

  • Mixed reality applications can address the limitations of traditional remote communication tools by better integrating users' surrounding physical environments.
  • As future collaborative work increasingly involves participants from diverse locations, mixed reality technologies that enable seamless transitions and collaboration across heterogeneous spaces will be critical.

• Research Motivation and Related Work:

  • Existing methods attempt to align environments through physical or nonlinear transformations. However, both approaches introduce distortions and scaling issues that negatively impact user experience.
  • Exploring localized alignment methods—aligning only the spatial areas relevant to the current task and adjusting alignment when users transition to new surfaces—offers a more flexible solution.

Solution

• Methodology and Solution:

  • Proposed a new concept called "Partially Blended Realities (PBR)" to support remote user collaboration across dissimilar spaces through localized alignment design.
  • Introduced two solutions:
    1. Realignment: Users actively trigger surface alignment state transitions, dynamically adjusting the environment through gradual animations or instant switches.
    2. Overlay: Displays two layers of remote users and surfaces simultaneously, allowing users to achieve indirect alignment by focusing on different layers.
  • Developed a prototype system, "RealityBlender," to test and validate the localized alignment techniques.

• Innovations:

  • Localized alignment focuses only on task-relevant physical surfaces rather than attempting to align the entire space.
  • The proposed alignment techniques allow users to transition within localized areas and switch spatial reference points actively or passively.
  • The design supports dynamic use of spaces (e.g., transitioning between a desk and a whiteboard), enabling users to freely adjust alignment modes during collaboration.

• Implementation Steps and Key Technologies:

  • Implementation of the "RealityBlender" System:
    • Developed using Meta Quest 2 and Unity3D, integrating the Oculus XR Toolkit and Meta Avatars SDK to support user interaction and multi-user networking.
    • Users can define physical surfaces via gestures and virtualize these surfaces as collaborative action spaces.
    • Provided dynamic configuration tools to match and align remote and local surfaces.

Research Findings

• Key Findings:

  • User studies revealed that localized surface alignment effectively enhanced the sense of "being together" during collaboration.
  • Triggering realignment after users reached the target location caused some discomfort, while preemptive alignment resulted in smoother transitions.
  • The overlay technique allowed users to switch focus implicitly, though displaying multiple users and surfaces simultaneously caused some confusion.

• Advantages and Comparisons:

  • Compared to existing global alignment methods, PBR directly addresses coordination issues in heterogeneous spatial environments, avoiding discomfort caused by distorted spatial scaling.
  • Demonstrated better scalability in supporting dynamic multi-surface collaboration.

• Experimental and Evaluation Results:

  • User studies comparing three alignment conditions (realignment, overlay, and physical consistency alignment) showed that localized alignment techniques outperformed global methods in flexibility for cross-space operations.
  • Designs such as "real-time triggering" and "surface-level switching" significantly improved the collaborative experience.
  • Users gradually adapted to the overlay technique but suggested visual fading improvements for transition layers.

• Limitations and Future Directions:

  • All experiments were conducted in simplified physical setups with controlled operations, which may not fully reflect the complexity of real-world office environments.
  • Localized alignment techniques may become more complex when handling multi-user, multi-surface scenarios, requiring further optimization to support larger-scale collaboration.
  • Future work could explore extending "Partially Blended Realities" to support more complex dynamic configurations, including repositioning, scaling, and rotating virtual surfaces to accommodate broader collaboration models.

Conclusion

This paper introduces the concept of "Partially Blended Realities" and its prototype system, addressing the alignment challenges of heterogeneous spaces in remote collaboration through localized alignment techniques. The study demonstrates a flexible and scalable approach to collaboration, with future research opportunities in exploring more complex mixed reality scenarios and user interfaces to expand the possibilities for remote mixed reality work.