Impression Evaluation of Interaction Among Robots based on Collision Avoidance Behaviors

Abstract

For achieving effective human-robot interaction, robot behaviors must be designed that humans can easily predict and understand. Humanoid robots are equipped with such mechanisms as facial expressions or gestures for communicating their internal states. Although humans can easily grasp such facial expressions or gestures, unfortunately, non-humanoid robots lack such mechanisms. In this research, we employ robot behaviors and their interactions with the other robots for expressing intentions and emotions considering situations where multiple non-humanoid robots collaboratively perform tasks. Here, we focus on such interaction behaviors among robots as collision avoidance and evaluate the effects of the differences of such behaviors on human impressions of the robots. Our experimental results of the subjective evaluations showed that the human impressions of the robots were affected by not only their speed but also by their interactions, including the collision avoidance behaviors of the robots.