Team-Centered Virtual Interactive Presence for Adjustable Autonomy

Abstract

Current solutions for autonomy plan creation, monitoring and modification are complex, resulting in loss of flexibility and safety. The size of ground control operations and the number of aviation accidents involving automation are clear indicators of this problem. The solution is for humans to naturally collaborate with autonomous systems. Visual spatial information is a common reference for increased team situation awareness between humans and adaptive autonomous systems such as robots, satellites, and agents. We use a teamcentered, augmented reality spatial dialog approach to improve human-automation interaction for collaborative assembly, repair, and exploration. With our spatial dialog approach, we provide a data-rich virtual interactive presence for collaborating with autonomous systems by incorporating the spatial context along with the spoken context.Spatial Dialog is an interaction and communication technology that uses a deeper understanding of spatial context and a richer spatial vocabulary to realize improved human-computer symbiosis. Spatial context is useful for communicating, understanding and remembering information. It includes the location of objects and people, and the place in which events are occurring. Using computers that see, we can pick out faces in a scene, track people, and follow where people are pointing and the objects they use. Using augmented reality and a real-time understanding of the spatial scene, we can overlay information on the real world or virtual models as a key spatial component of the dialog process, especially for remote virtual tele-presence, tele-supervision, and tele-science cases.