Construction of Cognitive Environment by Joint Attention Behaviors Based on Relevance Theory

Abstract

In recent years, various kinds of communication robots are active in our lives. However, most of them consider only the existence of the speaker and the listener. In this research, we aim to realize smoother communication by constructing the relationship between the speaker, the listener and the object existing in the environment, and we have constructed a system that the robot guesses the object that the human is referring to. We derive the degree of relevance to objects existing in the environment from two perspectives of verbal information based on relevance theory and non-verbal information based on joint attention behaviors, and actually use the prepared scenario to interact with the communication robot. In the interaction, it was confirmed that the object actually referenced by human being matched with the object obtained by this system.