Effect of Shared-attention on Human-Robot Communication

In our pursuit of ways to quantitatively evaluate communication between humans and robots, we recently focused on the effect of shared attention on human decision-making. We used a head robot that can make facial expressions and has human face tracking capability, and designed the interaction so that the robot often looked at the same objects the subjects looked at. Subjects were asked to select a color name from two candidates while the robot made a recommendation. The ratio that subjects selected the recommended candidate was used as a measure of the robot’s persuasive communication ability. We found that the matching ratio was correlated with the period time that shared attention (SA) was established for a group of submissive subjects. There was a significant difference in the matching ratio between a high-SA group and a low-SA group. This suggests that establishing shared attention is important for human-robot communication.