A quadrotor platform for bio-inspired navigation experiments

Due to developments driven mostly by the demands of the mobile communication industry and preparatory work by model aircraft flyers, it is possible to construct different sorts of electrically powered flying robots that can carry an interesting amount of payload in sensors and computers. We are interested in realising autonomous navigation capabilities on real flying robots which should suffice the following requirements: • self-sufficient in the sense of not relying on external systems of reference such as GPS, visual tracking or similar means • self-sufficient in the sense of being able to do all necessary computations onboard the flying robot • usable indoors and outdoors Specifically we are working with quadrotor helicopters as the experimental platform since they can be built robust and powerful enough for indoor and outdoor flight, besides having other useful characteristics of locomotion (near holonomicity). Since the problems listed above are not easily soluble in general and conventional methods are easily subject to brittleness we expect bio-inspired solutions of parts of those problems to be the most promising path. These solutions pertain to sensory modalities, sensori-motor coupling and cognitive models built on top of the former two [2], [9]. At the Cognitive Robotics Group at the Department of Computer Science, Humboldt-Universität zu Berlin, this project has been going on for about three years ([4]).