This work presents DreamLander, a model-based reinforcement learning algorithm built on DreamerV3, designed to generate optimal guidance and control for a 6-degree-of-freedom entry descent landing problem.

This work develops a generalized reward function for multi-agent reinforcement learning (MARL) based spacecraft inspection. Unlike prior methods tied to fixed inspection points, trained agents have full control over image capture timing and location, informed by 3D reconstruction analysis of orbital objects.

This study evaluates D-NeRF, TiNeuVox, and HexPlane for reconstructing tumbling Resident Space Objects (RSOs) under high-fidelity orbital lighting conditions.