Vision based navigation for autonomous space exploration

As a passive light-weight system with a natural adaptation capacity to the environment, mimicking the human capacity for detecting hazards, vision-based navigation for space applications has been the subject of a sustained research effort in Europe for more than ten years. The “Navigation for Planetary Approach and Landing” (NPAL) ESA/Astrium Satellites project paved the way for a new European autonomous vision based navigation system called VisNAV, aiming at a proof of concept, with an elegant breadboard realization, preparing for the next step to flight demonstration. This paper presents the consolidated design, and HW/SW architecture of the real time implementation. It also addresses the validation strategy from simulation, making extensive use of virtual scene generation through a realistic modeling environment, to in-flight demonstration experiment, as well as the achievable performances.