On behalf of the European Space Agency (ESA), EFS and the Institute for Mechatronics and Robotics of the German Aerospace Centre (DLR) are collaborating for the project DeLeMIS (Demonstration of Learning Machines for Improvements in Space Missions). The main objective is to investigate and demonstrate state-of-the-art approaches for self-learning systems aiming at improvements in planetary space missions, for instance, exploration of the moon surface.
The intended scenario considers autonomous navigation of a rover on an unknown terrain without human intervention. To this end, algorithms related to the latest AI research will be employed for environment perception and behaviour control. On the one hand, the rover is supposed to learn the ground characteristics of different areas and their boundaries, using camera inputs as well as through the wheel interaction with the ground. On the other hand, the system shall learn the best behaviour and the most proper strategy for its motion and navigation across each ground surface, while regarding the safety and robustness aspects.
Project DeLeMIS started in March 2022 with an intended duration of two years. By the end of the project, some of the investigated algorithms will be implemented and experimentally validated on a rover prototype developed by DLR. In the end, there will be a demonstration and the results will be presented at the facilities of the European Space Research and Technology Centre (ESTEC) in Noordwijk, Netherland.