Examining Physiological Response and Facial Expression as Indicators of Trust in a Robot Partner

Developments in the assistive capabilities of robots have allowed them to be more prevalent across domains from manufacturing to healthcare to education. To achieve more efficient and fluent collaborations between robots and humans, it has become necessary to explore how humans trust robot partners during collaborations. Current measurements of human trust in a robot partner typically rely on reflective, subjective survey measures, which are difficult to incorporate and update in the robot’s decision-making models for real-time deployment. To address this challenge, we aim to identify potential objective measures of trust, which can be incorporated into the robot’s decision-making process. We examine physiological data (Electrodermal Activity (EDA), Skin Temperature (TEMP), and Blood Volume Pulse (BVP)) and facial expression in an in-person ( \(n=30\) ) human–robot collaboration study and report the effect of trust on these objective measures. The results indicate that the participants who distrusted the robot partner experienced higher EDA, higher TEMP, and lower BVP variations than participants who trusted the robot. For facial expression, participants who distrusted the robot had more intense facial expressions than participants who trusted the robot. Ultimately, these findings can be used to further research on developing a real-time model of human trust in robots.