Human-machine interaction as a model of machine-machine interaction: how to make machines interact as humans do

Turn-taking is one of the main features of communicative systems. In particular, it is one of the bases allowing robust interactions in imitation, thanks to its two linked aspects, i.e., communication and learning. In this article, we propose a simple model based on the interaction of two neural oscillators inhibiting each other which explain how 'turn-taking' may emerge dynamically between two agents. An implementation of the model on a simple robotic platform made of one CCD camera and one simple arm (ADRIANA platform) is detailed. Results showing the emergence of a 'turn-taking' dynamics on this platform are discussed and an extension in simulation for a larger scale of parameters in order to validate robustness is given.