Pretrained Approximators for Low-Thrust Trajectory Cost and Reachability

2026-05-26

AI Summary

This work addresses the high computational cost of optimal control solvers typically required for evaluating fuel consumption and accessibility in low-thrust orbit transfers. To overcome this limitation, the authors propose a neural network–based surrogate model trained on a large-scale dataset generated via a homotopy ray strategy. By incorporating scaling laws inherent to low-thrust trajectories, the method introduces a self-similar coordinate transformation that enables cross-domain generalization across varying semi-major axes, inclinations, and central bodies—eliminating the need for retraining. The approach demonstrates high accuracy and efficiency in predicting performance for both single- and multi-revolution transfers, with validation on public datasets as well as complex missions involving multiple asteroid flybys and rendezvous. The implementation code and associated data are publicly released.

Abstract

Low-thrust trajectory design relies heavily on repeated evaluations of fuel consumption and transfer feasibility, which require expensive optimal control solutions. In this work, we show these quantities can be accurately approximated by machine learning surrogates, enabling fast and scalable evaluation across a wide range of scenarios. By increasing both dataset size and model capacity, we observe that low-thrust trajectory optimization follows a scaling law, with performance improving linearly with the logarithm of training data and network parameters, and no evidence of saturation within the explored regime. Guided by this observation, we construct a large-scale dataset using the proposed homotopy-ray strategy tailored to mission design requirements. A key is the introduction of a self-similar transformation, which allows generalization across semi-major axes, inclinations, and central bodies avoiding retraining. As a result, the same neural approximator can be applied to diverse orbital environments and mission classes. The proposed models accurately predict optimal fuel consumption and minimum transfer time for single- and multi-revolution transfers. Their performance and generalization are demonstrated on a public dataset, a multi-asteroid flyby problem from the Global Trajectory Optimization Competition, and an asteroid rendezvous mission design. The models and datasets are released as open-source to support the space community.

Problem

• low-thrust trajectory
• fuel consumption
• transfer feasibility
• optimal control
• trajectory optimization

Innovation

• low-thrust trajectory
• machine learning surrogate
• scaling law
• self-similar transformation
• orbit transfer optimization