Spatial movement concepts in human-robot-shared environments

Annika Peters, Petra Weiß, Marc Hanheide Robots are more and more leaving their autonomous working areas to move into environments they share with human users. Picking up ideas of spatial cognition in shared environments, the project will target at movement and navigation concepts affected by the continuously perceived situational context in a home tour setting. Looking into recent developments in robotics, today's mobile domestic robots are able to navigate autonomously in spatial environments like rooms; they can follow humans and avoid colliding with obstacles. So far, this is achieved usually by goal-directed point to point navigation. However, implementing following behavior alone is not sufficient to achieve natural and realistic HRI. There are situational constraints like for instance small spaces or multiple persons in the robot’s vicinity prohibiting the classical way of user-guided navigation. In this context, the project aims not only at a more situation-adequate guidance of the robot by the user but also at the possibility of mutual guidance by robot and user to allow for joint exploration as a means to facilitate learning by implicit interaction exploiting spatial concepts. The objective of this project is to provide a mobile robot with semantically rich and intelligent spatial concepts to allow spatial interaction with humans considering relevant aspects of a particular situation, on other words to provide a robot with situation awareness and concepts of spatial intelligence. The project will concentrate primarily on the aspect of implicit communication by movement in space: How can robot and human user interact by indicating intentions by movement, positioning, and orientation? To reach this goal the project envisions an iterative approach comprising implementation and evaluation. Starting off from analysis of existing corpora of HRI studies on joint exploration, novel movement concepts can be developed on the basis of existing reactive and metric robot navigation. Expected results concern particularly a yet unachieved level of spatial intelligence in domestic environments, a set of novel and semantically rich spatial movement concepts on the basis of situational contexts, a robotic system that exhibits and understands non-verbal, movement-driven communication patterns as well as insights regarding differences between human-human joint spatial movement and human-robot movements.[view:groupmembers==212]