Social Robot Navigation adapted to Time-dependent Affordance Spaces: a Use Case for Caregiving Centers

摘要

The use of socially assistive robots in real environments, such as nursing homes or geriatric residences, is spreading in recent years. Social robot navigation in these complex environments, with multiple users and dynamic objects, is an essential task for the next generation of service robots. This navigation must respect social rules, for example not to interrupt an interaction between two people or between a person and an object. Current navigation frameworks in robotics literature do not take into account the social complexity of the scenarios like, for example, the relation of the objects and their use by people with time. This article presents an approach to the idea of time-dependent social mapping. The main novelty is the definition and development of time-dependent affordance spaces. Each object has a zone that vary in function of time and the activities scheduled by the center's staff. Therefore, the planning of the best path by the robot takes into consideration this variation on time achieving a higher degree of social navigation. Several use cases have been performed in a simulated scenario to asses the robustness and validity of the proposal using these temporal variables.