A social practice oriented signs detection for human-humanoid interaction

In this work we propose a cognitive architecture, based on the Social Practice (SP) theory, aimed at the modeling of socially adaptive robots, able to interact with people, recognizing and interpreting the specific social context where it is acting. The proposed social robot is able to recognize and interpret social signs during ongoing social practices. The cognitive architecture is inspired by the well-known Psi model, and it is equipped with a Social Practice Engine that manages the whole conduct of the robot. The use of such an architecture simplifies and makes more natural the interaction between human beings and a robot. In particular, the scenario of a robot fulfilling the tasks of welcoming people in an office environment is being considered.