Which is Warmer, the Cake or the Oven? Unlocking Thermal Conductivity for Virtual Reality Interaction

Paper Title

Which is Warmer, the Cake or the Oven? Unlocking Thermal Conductivity for Virtual Reality Interaction

Publication Info

• Topic area: Thermal haptics in virtual reality (VR) for material discrimination.
• Keywords: Thermal feedback, virtual reality, haptic interaction, thermal conductivity, material discrimination, multimodal interaction, user experience, WEART gloves, immersive systems, thermal rendering.

Background and Problem

• Problem / challenge: Existing VR systems primarily use static temperature rendering, which fails to capture the dynamic thermal cues necessary for realistic material discrimination. The interplay of visual and thermal cues in VR remains underexplored, particularly under conditions of sensory incongruence.
• Significance: Dynamic thermal feedback can enhance realism, immersion, and material perception in VR, supporting applications in training, telepresence, and entertainment. However, current systems lack a robust framework for rendering material-specific thermal dynamics.
• Motivation and related work: Prior research has demonstrated the potential of thermal feedback for enhancing presence and material judgments in VR. However, these studies often focus on static or isolated thermal cues, neglecting the role of dynamic heat transfer and its integration with multimodal feedback. This paper builds on advances in thermal displays and wearable devices to explore conductivity-based thermal rendering.

Solution

• Proposed approach: A physics-informed conductivity model that simulates dynamic heating and cooling profiles based on material properties (metal, glass, wood) for VR interaction.
• Novelty:
  1. Development and validation of a conductivity-based thermal rendering model for VR.
  2. Empirical demonstration of how dynamic thermal cues enable perceptually reliable material discrimination.
  3. Insights into the role of visual-thermal congruence and user expertise in interpreting thermal feedback.
  4. Device-agnostic framework for integrating thermal cues into multimodal VR systems.
• Procedure and key techniques:
  • Participants interacted with virtual objects (metal, glass, wood) using WEART TouchDIVER Pro gloves and a Meta Quest 3 headset.
  • Thermal feedback was rendered using a simplified heat transfer model, with material-specific conductivity values (metal = 2.0, glass = 1.1, wood = 0.7).
  • Three experimental conditions were tested: visual-thermal congruence, thermal-only, and visual-thermal incongruence.
  • Participants were divided into informed and non-informed groups to assess the impact of prior knowledge on performance.
  • Task performance, subjective experience (HX ratings), and physiological baselines were analyzed.

Results

• Concrete findings:
  • Sorting accuracy was highest under visual-thermal congruence (71%), lower under thermal-only (56%), and lowest under visual-thermal incongruence (44%).
  • Thermal feedback ratings followed real-world conductivity order (Metal > Glass > Wood), validating the model.
  • Informed participants outperformed non-informed participants in accuracy, especially under thermal-only conditions (e.g., 75% vs. 40% accuracy for metal).
• Advantage over baselines:
  • Dynamic thermal cues enabled material discrimination even in the absence of visual information, outperforming static thermal rendering approaches.
  • Informed participants demonstrated better alignment between subjective ease ratings and objective accuracy, indicating the value of onboarding.
• Experiments / evaluation:
  • Conducted with 38 participants (balanced by gender, mean age = 30.14 years).
  • Tasks involved sorting virtual objects under three visual-thermal conditions, with accuracy, sorting time, and HX ratings as metrics.
  • Validation of thermal rendering was performed using a FLIR thermal camera to compare theoretical and rendered temperature profiles.
• Limitations and future work:
  • Hardware constraints limited the absolute fidelity of thermal cues, particularly for wood and glass.
  • Short-duration tasks in controlled settings may not generalize to longer or more complex interactions.
  • Participant pool was relatively homogeneous, and individual differences in thermal sensitivity were not fully explored.
  • Future work should test the model across different devices, integrate multimodal feedback, and explore adaptive rendering for diverse user groups.

Summary

This paper introduces a conductivity-based thermal rendering model for VR that simulates dynamic heating and cooling profiles to enable material discrimination. The model was validated through experiments showing that thermal cues aligned with real-world conductivity properties and enhanced realism and immersion. Informed participants outperformed non-informed participants, highlighting the importance of onboarding in interpreting novel thermal feedback. While current hardware limitations constrain fidelity, the model provides a scalable and device-agnostic framework for advancing thermal haptics in VR. Future research should focus on multimodal integration, adaptive rendering, and broader user testing to fully realize the potential of conductivity-based thermal feedback.