Tight Constraint Prediction of Six-Degree-of-Freedom Transformer-based Powered Descent Guidance

TL;DR

The paper tackles real-time onboard generation of fuel-optimal 6-DoF powered descent trajectories. It extends Transformer-based tight-constraint prediction to the successive convexification framework, enabling reduced subproblem sizes and warm-starts via learned active constraint sets and initial guesses. Rotation-invariant data augmentation reduces the training burden while preserving feasibility, and the approach demonstrates substantial runtime reductions on a Mars landing test case. The work shows that T-SCvx can deliver reliable, onboard trajectory optimization with real-time capabilities, potentially enabling higher safety margins and more challenging landing sites.

Abstract

This work introduces Transformer-based Successive Convexification (T-SCvx), an extension of Transformer-based Powered Descent Guidance (T-PDG), generalizable for efficient six-degree-of-freedom (DoF) fuel-optimal powered descent trajectory generation. Our approach significantly enhances the sample efficiency and solution quality for nonconvex-powered descent guidance by employing a rotation invariant transformation of the sampled dataset. T-PDG was previously applied to the 3-DoF minimum fuel powered descent guidance problem, improving solution times by up to an order of magnitude compared to lossless convexification (LCvx). By learning to predict the set of tight or active constraints at the optimal control problem's solution, Transformer-based Successive Convexification (T-SCvx) creates the minimal reduced-size problem initialized with only the tight constraints, then uses the solution of this reduced problem to warm-start the direct optimization solver. 6-DoF powered descent guidance is known to be challenging to solve quickly and reliably due to the nonlinear and non-convex nature of the problem, the discretization scheme heavily influencing solution validity, and reference trajectory initialization determining algorithm convergence or divergence. Our contributions in this work address these challenges by extending T-PDG to learn the set of tight constraints for the successive convexification (SCvx) formulation of the 6-DoF powered descent guidance problem. In addition to reducing the problem size, feasible and locally optimal reference trajectories are also learned to facilitate convergence from the initial guess. T-SCvx enables onboard computation of real-time guidance trajectories, demonstrated by a 6-DoF Mars powered landing application problem.

Figures (4)

• Figure 1: tscvx inputs include the initial state and iteration number. The output of the constraint neural network, $\tau(\theta^{(i)})$, is the set of tight constraints, and the outputs of the initial guess neural network, $t_f^*(\theta^{(i)})$, $x_*(\theta^{(i)})$, and $u^*(\theta^{(i)})$, are the optimal final time, state, and control.
• Figure 2: Box plots for tscvx and scvx showing mean, median, and standard deviations.
• Figure 3: 6-dof minimum fuel trajectory computed by tscvx with thrust vectors in red.
• Figure 4: Memory vs. inference time for each warm-start algorithm type.

Theorems & Definitions (3)

• Definition 1
• Definition 2
• proof