Supporting Human-Robot Teams in Social Dynamicism: an overview of the Metaphoric Inference Framework

Abstract

Metaphoric classification of social interaction in human-robot teams can provide a useful frame for directing robot-human engagement while assisting the robot in sense-making of dynamic world behaviors. This paper describe a framework built from principles of embodiment to show how members of teams are not separated from their world, but make sense and interact in a world via a continuous (rather than causal) flow of engagement facilitated by two forms of perception. In this context, we distinguish between direct and reflective perception, arguing that agents socially engage via these modalities of perception through body, language, and context. We then argue that these forms of perception can direct the use of metaphors. The metaphors, in turn, act as classification frames for robot social intelligence using established human schemas.