Filter Banks and Integrity Monitoring for Visual UAV Navigation

This paper outlines current and future navigation for small unmanned aircraft. An example scenario is the exploration of disaster areas where the aircraft flies at low altitudes and where a-priori hints like maps are outdated or otherwise not reliable. The paper tackles navigation problems where satellite data might be temporarily unavailable. It presents a state estimator which uses satellite, inertial, and visual data. In order to determine and correct errors, a filter bank is proposed to analyze the output of INS/GNSS integration and to exclude GNSS failures from further processing. The paper also presents results from flight tests with a small helicopter where a suitable trajectory can be calculated within incomplete navigation satellite constellations. Finally, the paper discusses some open issues for future work.