Modeling and Implementation of Intra-Hug Gestures During Dialogue for a Huggable Robot

The potential of social touch, a behavior that uses physical embodiment, has been actively studied for adaptation to social robots because of both physical and mental benefits. Social touch from robots, including hugs, reports positive effects similar to human-human interaction. Huggable robots have been developed to make intra-hug gestures such as patting and rubbing during hugs. However, no models have been built to determine the frequency, type, contact area, initiation timing, and duration of such gestures during dialogue that provide support. As a first step toward resolving this gap, this study aimed to build a model of intra-hug gestures during dialogue. We gathered data from participants acting as addressees with a mannequin as a speaker. From the probability distribution estimated from data analysis, we constructed a model to determine the parameters of the gestures. We implemented an autonomous dialogue system with intra-hug gestures through a robot. The findings of this study have implications for dialogue designs with intra-hug gestures for huggable robots.