Determining Perception Thresholds for Real and Virtual Inclinations While Cycling in Virtual Reality

Paper Title

Determining Perception Thresholds for Real and Virtual Inclinations While Cycling in Virtual Reality

Publication Info

• Topic area: Human perception in virtual reality cycling systems.
• Keywords: Virtual reality, cycling, perceptual thresholds, inclination mismatch, sensory integration, exergames, immersive training, vestibular feedback, proprioception, VR sickness.

Background and Problem

• Problem / challenge: Mismatches between real and virtual inclinations in VR cycling can disrupt immersion and induce cybersickness. Prior research lacks empirical thresholds for these mismatches.
• Significance: Understanding perceptual thresholds can expand VR design possibilities for immersive training, exergames, and rehabilitation systems.
• Motivation and related work: Previous studies on visuo-haptic mismatches and VR illusions have demonstrated the dominance of vision over other senses. However, the specific thresholds for inclination mismatches in VR cycling remain unexplored.

Solution

• Proposed approach: Empirical measurement of perceptual detection thresholds for mismatches between real and virtual inclinations during VR cycling.
• Novelty:
  1. Quantification of congruence windows for real-to-virtual inclination mismatches.
  2. Identification of asymmetric tolerance for uphill vs. downhill mismatches.
  3. Integration of visual, vestibular, and proprioceptive cues in a VR cycling simulation.
  4. Development of a staircase procedure to measure perceptual thresholds.
• Procedure and key techniques:
  • Participants cycled on a tilting stationary bike in a VR city simulation.
  • Real inclinations ranged from -15% to +15%, while virtual inclinations extended up to ±50%.
  • A multiple staircase procedure measured upper and lower perceptual thresholds for each real inclination.
  • Participants judged whether virtual slopes appeared "too steep" or "too flat," using handlebar buttons for binary responses.

Results

• Concrete findings:
  • Virtual inclinations can deviate significantly from real inclinations without detection: e.g., real -15% downhill paired with virtual -50%, and real +15% uphill paired with virtual +31%.
  • Thresholds are asymmetric, with greater tolerance for downhill mismatches.
  • No significant correlations between thresholds and demographics, VR/cycling experience, or subjective load.
• Advantage over baselines: Provides actionable thresholds for manipulating virtual inclinations without disrupting immersion, expanding the design space for VR cycling systems.
• Experiments / evaluation:
  • 30 participants (ages 21–56) completed counterbalanced upper and lower threshold measurements.
  • Metrics included perceptual thresholds, posture analysis, and standardized questionnaires (Borg-RPE, NASA-TLX, VRSQ, IPQ).
  • Results showed stable thresholds across demographic and experiential variables.
• Limitations and future work:
  • Sample biased toward younger participants; findings may not generalize to older populations.
  • Thresholds measured at P50 detection levels; tolerance thresholds (P75) remain unexplored.
  • Future studies should investigate continuous slope changes and broader perceptual manipulations (e.g., pedaling resistance, speed).

Summary

This study establishes perceptual detection thresholds for mismatches between real and virtual inclinations during VR cycling, revealing congruence windows within which deviations remain unnoticed. Results show asymmetric tolerance for uphill and downhill mismatches, with greater flexibility for virtual downhill exaggerations. The findings enable designers to manipulate inclinations in VR cycling systems for immersive training, exergames, and rehabilitation, while maintaining user immersion and comfort. Future work should refine tolerance thresholds, explore long-term effects, and extend findings to other VR domains.