Modeling Social Robot Navigation: From Human Observation to Proxemics-Based Scenario Simulation
摘要
Proxemics, how individuals perceive physical space around them, is essential for designing social robot navigation. This study explored human-robot proxemics (HRP) by looking at personal space as a continuous shape around a person. Drawing from real-world observations of human-human interactions, we identified six situations that reflect natural navigation patterns relevant to human-robot interactions. In parallel, using experimental data reported in the literature, we developed spatial proxemics models that represent the personal space surrounding individuals. We then integrated these models into a set of human-robot interaction scenarios within a simulation environment that also incorporates a robot’s navigation behavior, to study more realistic investigations of proxemic dynamics. The simulation outcomes were assessed for navigation behaviors that appeared uncomfortable, revealing two uncomfortable HRI scenarios. This methodology enables structured explorations of robot violations of proxemic patterns, laying the groundwork for future real-robot experiments that examine the spatial structure of proxemics in human-robot interaction scenarios. (R. Burdman and E. Nahum—These two authors contributed equally to this work).