Dynamic proxemia modeling approach (DPMA) for navigation and interaction with a group of persons

Abstract

Most of approaches in Human-Robot Interaction consider a static proxemia independently from the F-Formation type of the people with which we interact and the support of communication (speech, phone, ...). In this paper, we propose a new approach by formalizing dynamic proxemia depending on both the group of people F-Formation and on the activity to accomplish. We develop "Dynamic Proxemia modelling approach" (DPMA) by considering different criteria such as social conventions and the activity model to produce a heat map of social interaction and navigation. We develop experiments using ROS and Stage for simulation of a guide robot in a mediation event example showing promising results of our approach.