Investigating Resolution Strategies for Workspace-Occlusion in Augmented Virtuality

TL;DR

Augmented Virtuality enables embedding physical content inside virtual spaces, but misalignment during virtual locomotion causes occlusion of physical objects by virtual elements. The authors compare two automatic resolution strategies, Redirected Walking (RDW) and Automatic Teleport Rotation (ATR), in a user study set in a virtual forest, measuring efficacy, cybersickness, orientation, workload, agency, and user satisfaction. ATR resolves occlusion more reliably but shows potential drawbacks in cybersickness and user acceptance, whereas RDW provides a quieter, less perceptible intervention with somewhat lower occlusion resolution. The results inform design choices for AV systems by balancing occlusion resolution against user comfort and orientation, suggesting context-driven deployment of ATR or RDW. Overall, both strategies effectively mitigate unwanted occlusion, with tradeoffs that depend on task demands and user preferences.

Abstract

Augmented Virtuality integrates physical content into virtual environments, but the occlusion of physical by virtual content is a challenge. This unwanted occlusion may disrupt user interactions with physical devices and compromise safety and usability. This paper investigates two resolution strategies to address this issue: Redirected Walking, which subtly adjusts the user's movement to maintain physical-virtual alignment, and Automatic Teleport Rotation, which realigns the virtual environment during travel. A user study set in a virtual forest demonstrates that both methods effectively reduce occlusion. While in our testbed, Automatic Teleport Rotation achieves higher occlusion resolution, it is suspected to increase cybersickness compared to the less intrusive Redirected Walking approach.

Figures (6)

• Figure 1: RDW strategy: The desk is initially occluded by a virtual tree (left). During the main task, which involves physical walking around other trees, subtle gains are applied to redirect the user (middle). By the end of the main task, the occlusion is resolved (right).
• Figure 2: ATR strategy: The left image shows a teleport preview without occlusion, requiring no adjustment. In the right image, the user moved the preview so that the tree would now occlude the desk, prompting the system to automatically rotate the preview upon teleportation.
• Figure 3: The possible manipulations of the desk to find an occlusion-free position (orange). The translation is always directed towards and the rotations around the origin (blue cross). With the ATR strategy (a), the origin is the center of the teleport-preview's position. With the RDW strategy (b), the origin is the user's position.
• Figure 4: Formular to calculate the updated position ($p'$) and yaw rotation ($\theta'$) with the RDW strategy. The signs of the rotation and curvature gains indicate whether the desk should be rotated clockwise or counterclockwise around the user.
• Figure 5: The sum of squared distances between points and their cluster centers (WCSS) of each number of clusters (k). The k-value, where the WCSS starts to decrease much slower, indicates an optimal clustering. In this case, the plot indicates an optimal clustering with either k=3 or k=4 for Preference and k=3 for Usability and User Experience.