Researchers at The University of Luxembourg Develop a Method to Learn Grasping Objects on the Moon from 3D Octree Observations with Deep Reinforcement Learning

The goal of planetary exploration is to improve science by revealing new information about the geology and resource potential of other worlds. Extraterrestrial robotic systems are crucial in their acquisition for either in-situ analysis or sample return, which is a significant component of this project’s sample analysis. In order to extract hydrogen and oxygen for the local generation of rocket fuel and the functioning of life support systems, reports of these discoveries are also required for future in-situ resource usage. …

3d learning machinelearningnews moon objects reinforcement reinforcement learning researchers university