Human-aware Robot Navigation based on Time-dependent Social Interaction Spaces: a use case for assistive robotics

Abstract

Social navigation in care-giving environments - nursing homes, geriatric residences, for example - is an essential task for future generations of service robots. Navigating in a complex environment with elderly people, clinical staff or companions implies the need to adapt to social conventions the planned routes, the velocity of approach to people and the spaces of interaction between robot-human, human-human and human-object. Currently, navigation algorithms do not take into account the social complexity of the scenarios like, for example, their relationship with the hour of the day or the activities carried out in these scenarios. This article presents a first approach to the idea of time-dependent social mapping, where the planning of the social route by the robot takes into consideration variables that depend on the time and schedules of use of certain spaces. To this end, this article describes how the spaces of human-object interaction vary as a function of time and how this affects the social navigation planned by the robot. Several use cases have been performed in a simulated environment to asses the improvements in the robot social navigation using these temporal variables.