Compositional Diffusion Models for Powered Descent Trajectory Generation with Flexible Constraints

This work introduces TrajDiffuser, a compositional diffusion-based flexible and concurrent trajectory generator for 6 degrees of freedom powered descent guidance. TrajDiffuser is a statistical model that learns the multi-modal distributions of a dataset of simulated optimal trajectories, each subject to only one or a few constraints that may vary for different trajectories. During inference, trajectories are generated simultaneously over time, providing stable long-horizon planning, and constraints can be composed together, increasing the model's generalizability and decreasing the training data required. The generated trajectories enable efficient optimizer initialization, increasing its robustness and speed.