Integrating the physical environment into mobile remote collaboration

We describe a framework and prototype implementation for unobtrusive mobile remote collaboration on tasks that involve the physical environment. Our system uses the Augmented Reality paradigm and model-free, markerless visual tracking to facilitate decoupled, live updated views of the environment and world-stabilized annotations while supporting a moving camera and unknown, unprepared environments. In order to evaluate our concept and prototype, we conducted a user study with 48 participants in which a remote expert instructed a local user to operate a mock-up airplane cockpit. Users performed significantly better with our prototype (40.8 tasks completed on average) as well as with static annotations (37.3) than without annotations (28.9). 79% of the users preferred our prototype despite noticeably imperfect tracking.