Autonomous localization and control of airplanes has been proven to be feasible. On Earth, such systems usually use inertial measurement units, altimeters and GPS for localization and vehicle stabilization. During the last decade, such sensors have been miniaturized and their accuracy has been improved. This technology opens a large range of applications and therefore can be integrated in small flying devices.

For planetary exploration, the GPS is not available and many problems arise for long-range missions. The use of proprioceptive sensors is not sufficient because they lead to unbounded error propagation. The position information must be completed with exterioceptive sensors such as range sensors and vision. In our application, vision will be used to track a predefined targets or contours, and localize the airplane in relation with the skyline.

The three main work directions for navigation are:

• Development of ligthweight and low power consumption Hardware

• Realization of a dynamic model of the airplane

• Design of a model-based controller and a navigation algorithm