Analysis of Thermal Properties of Artificial Fingers for Recording Thermal Transients During Hand-Object Interactions

Abstract

Haptic feedback is a key component in human-computer interaction, providing human users with immersive and realistic experiences. Thermal feedback, a form of haptic feedback, enables users to recognize the materials of virtual objects by simulating changes in skin temperature during hand-object interactions. In order to carry out these simulations, it is essential to characterize skin temperature changes as the basis for the simulations. One of the characterizing methods is utilizing the data on the change in skin temperature directly measured. Creating a dataset of the changes in skin temperature during hand-object interactions with human fingers presents challenges due to inconsistent measurement settings and laborious processes. An artificial finger with thermal properties similar to those of human fingers could overcome the problems. In the present study, we analyzed the tolerance range of the thermal effusivity for the artificial finger, aiming to replicate thermal transients akin to those in human fingers during hand-object interactions. This range was investigated using a thermal model that predicts skin temperature change based on the thermal properties of skin and materials. As a result, the thermal effusivity of the artificial finger is required to be in the range of 1090 to 1281 $(J/m^2$s$^{1/2}K)$ for the measurement to be valid when the contact material spans a wide thermal property range. Based on these findings, we discuss candidate materials for the artificial finger and the potential applications in a machine-learning model for thermal displays, as well as an automatic material classification system based on thermal cues for robot hands.