Landmark navigation with ATHENE

A brief overview on the work to develop a vehicle for visual landmark navigation is given. The intention is to enable an AGV (autonomous guided vehicle) to operate in surroundings without special cues or extra installed markers. For navigation, objects are being used that are also visual hints for the orientation of humans. The robot functions with a certain amount of a priori knowledge, like an existing map of the building and uses sensing to fill in the details in a corresponding map during a teach-in-ride. With the acquired knowledge and a given task the robot determines, what actions are necessary to reach its goal. This is achieved by a combination of modeling, planning, sensing and controlled actions. The results of navigational experiments with ATHENE are discussed.<>