Walking Through or Detour? Investigating Walking Paths with World‑Anchored Mixed Reality Objects

Paper Title

Walking Through or Detour? Investigating Walking Paths with World‑Anchored Mixed Reality Objects

Publication Info

• Topic area: Mixed reality (MR) walking behavior and collision avoidance
• Keywords: Mixed reality, walking paths, virtual obstacles, collision avoidance, dynamic effects, user behavior, head-mounted displays, detour behavior, walk-through rate, spatial interaction

Background and Problem

• Problem / challenge: Limited knowledge exists about how users navigate around or through virtual objects in MR environments, particularly regarding walking paths and factors influencing walk-through versus detour behavior. Prior research has focused mainly on detour routes and has not adequately analyzed walk-through scenarios or the effects of dynamic virtual object behaviors.
• Significance: Understanding walking behavior in MR environments is critical for designing safer and more effective MR systems, especially as MR becomes more integrated into daily life.
• Motivation and related work: Previous studies have explored collision risks, detour routes, and factors influencing walk-through rates but have largely ignored the impact of object shape, dynamic effects, and individual differences. This paper addresses these gaps by investigating walking paths for both static and dynamically disappearing virtual obstacles.

Solution

• Proposed approach: The study investigates walking behavior toward static and dynamic virtual obstacles in MR environments, focusing on factors influencing walk-through rates and walking paths.
• Novelty:
  1. Identification of factors influencing walk-through versus detour behavior for static virtual obstacles.
  2. Development of three dynamic effects (Fade, Wipe, Door) for virtual obstacles that disappear as users approach.
  3. Analysis of walking paths for static and dynamic virtual obstacles.
  4. Insights into designing safer MR environments with world-anchored virtual objects.
• Procedure and key techniques:
  1. Conducted two studies with 58 participants (28 for static obstacles, 30 for dynamic effects).
  2. Used HTC VIVE Focus Vision HMDs and motion trackers to record walking paths.
  3. Designed experiments with variables such as obstacle width, depth, opacity, type, and dynamic effect timing.
  4. Analyzed metrics including walk-through rate, lateral deviation, peak position, and subjective impressions (RPQ and CAQ).

Results

• Concrete findings:
  • Walk-through rates for static obstacles increased with obstacle width (17.1% for 0.4 m, 39.2% for 1.2 m) but were unaffected by depth, opacity, or height.
  • Dynamic effects reduced lateral deviation compared to static obstacles, with Wipe and Door outperforming Fade.
  • Most participants initially detoured around dynamic obstacles but transitioned to direct paths as the obstacles disappeared.
  • Walk-through rates varied widely among individuals, with some participants walking through obstacles in over 80% of trials.
• Advantage over baselines: Dynamic effects (Fade, Wipe, Door) significantly reduced lateral deviation and collision anxiety compared to static obstacles. Wipe and Door were particularly effective in minimizing detours.
• Experiments / evaluation:
  • Study 1: 4,032 trials analyzed for static obstacles with varying width, depth, opacity, and type.
  • Study 2: 3,120 trials analyzed for dynamic effects (Fade, Wipe, Door) with different timings and widths.
  • Metrics included walk-through rate, lateral deviation, peak position, RPQ, and CAQ scores.
• Limitations and future work:
  • Narrow participant demographics (university students) and limited device types (video see-through HMDs).
  • Restricted obstacle appearances and simplified walking scenarios (single obstacle, straight paths).
  • Future work should explore diverse environments, obstacle designs, and real-world contexts with pedestrians and obstacles.

Summary

This study investigated walking paths and walk-through behavior for static and dynamic virtual obstacles in MR environments. Key findings include the influence of obstacle width and individual differences on walk-through rates, the effectiveness of dynamic effects (Wipe and Door) in reducing detours, and the delayed detour timing caused by exposure to disappearing obstacles. These insights inform the design of safer MR systems by highlighting the need for personalized virtual object placement, appropriate dynamic effect timing, and regulatory guidelines for virtual object interactions.