Semi-Autonomous Control of Unmanned Ground Vehicles

Mobile Intelligence is building an open architecture hardware and software system that supports semi-autonomous control of ground robots using monocular vision. It will be open architecture, follow the JAUS/AS-4 standard, and be suitable for mounting on small robots, such as the Talon or PackBot. It will allow the operator to easily command the robot to perform actions such as “Go There”, “Follow Path”, or “Follow Me” and allow it to autonomously travel, freeing the operator to perform other tasks, including maintaining their situational awareness and safety. Three research thrusts will be followed. First, categorizing terrain traversability based on appearance and motion allows extrapolating nearby, highly accurate estimates into the far scene. Second, classifying objects based on their shape and texture allows estimating their size based on heuristics, and thus estimating their distance. Finally, learning scene affordances with respect to navigation will allow learning mappings of visual shape categories to navigation strategies. The information from these three perceptual processes will be aggregated into a symbolic map, and used by a planner to create routes through the environment.