Perceiving head, shoulders, eyes and toes in augmented virtuality interfaces for mobile robots

Occupancy grid maps derived from laser range finder readings can be used to create virtual worlds for planar environments. These virtual worlds can be augmented with video to create augmented virtuality displays. Such displays, suitably adapted to represent local terrain and pan-tilt-zoom cameras, can make it easier for operators to simultaneously perceive the pose of the robot and its relation to information in the world. We motivate this claim using two abstractions from a model of cognitive information processing: working memory and mental models. We then describe visualization techniques that try to minimize cognitive workload by displaying information about the "head", "shoulders", "eyes", and "toes" of a teleoperated robot.