Investigating and Predicting Impacts of Thermal Feedback on Human Sensation and Emotion

Recent advancements in thermal device technology present new possibilities for conveying sensations and emotions in diverse applications. Yet, systematic investigations into the design parameters of thermal feedback and their effects on elicited sensations and emotions remain underexplored. To address this gap, we created 36 thermal feedback patterns by systematically varying the amplitude of change, rate of change, and indoor temperature, and applied them across three body sites on the hand, forearm, and upper arm. We then collected perceived intensity, valence, and arousal ratings from 12 participants. The results revealed the efficacy of the thermal design parameters. For example, the amplitude of change served as the primary parameter influencing the intensity and arousal ratings across body sites, while it affected valence only when thermal feedback was applied to the forearm and upper arm. Using the collected data, we developed a neural network to predict the intensity sensation and emotion ratings elicited by thermal feedback. Our model outperformed three baseline machine learning models and demonstrated strong alignment with non-linear sensory and emotional responses. We present four guidelines for designing thermal feedback and discuss implications for future research and applications in haptic design.