Conceptual
Browse Instruct Curate

Transformer-Based Successive Convexification for 6-DoF Powered Descent Guidance

Generate fuel-optimal six-degree-of-freedom rocket powered-descent trajectories fast and reliably by training a transformer to predict, from problem parameters, the tight (active) constraint set and a feasible reference trajectory at the optimum. Use those predictions to form a minimal reduced-size problem and warm-start a successive-convexification solver, enabling real-time onboard guidance, and evaluate the approach on a 6-DoF Mars powered-landing problem.