Implementation of context-adaptive physical imitation between humans and robots

Abstract

In this paper we propose a model of context-adaptive imitation between multi-human and multi-robot. In this model, first, a humanoid robot extracts human's intention from his/her behaviors, utterances, and situations. Secondly, the robot communicates this intention not only to other robots through a network, but also to humans through the robot's body expressions. Thirdly, the robot who received the intention generates context-adaptive behaviors according to the situation. We implemented this robot system for a guided task for visitors using the architecture of situated modules. We discuss the validity and availability of the proposed model using the case-study of the guided task by the implemented robot system.