Development, validation and test of optical based algorithms for autonomous planetary landing

In this paper, a suite of tools and algorithms devoted to optical navigation for autonomous landing on planets and small bodies, currently under development at PoliMi-DAER, is presented. An hazard detection system based on a single camera and artificial neural networks selects the most suitable landing site for the lander. Autonomous guidance on board computes the trajectory to reach for the designated landing site, taking into account the lander dynamics and path constraints. The two systems are linked by vision-based navigation, which reconstructs the relative spacecraft states with respect to the ground, in particular by means of a camera through feature tracking and other typical sensors, such as Inertial Measurement Unit and altimeter. The ultimate goal is to create a whole Adaptive Guidance, Navigation and Control chain for autonomous landings. An experimental facility is currently under construction at PoliMi premises to verify, validate and test the aforementioned optical landing tools.