Diffusion Models for Generating Ballistic Spacecraft Trajectories

TL;DR

This paper introduces a novel application of score-based diffusion models to generate ballistic spacecraft trajectories from Earth to Mars, demonstrating their ability to learn and produce realistic transfer paths.

Contribution

It presents a new generative framework using diffusion models for space trajectory design and provides a comprehensive analysis and benchmarking of their performance in this domain.

Findings

Diffusion models can effectively generate realistic ballistic transfer trajectories.

The model's performance varies with size and temporal resolution, as shown by ablation studies.

Benchmark results indicate diffusion models' potential for trajectory design applications.

Abstract

Generative modeling has drawn much attention in creative and scientific data generation tasks. Score-based Diffusion Models, a type of generative model that iteratively learns to denoise data, have shown state-of-the-art results on tasks such as image generation, multivariate time series forecasting, and robotic trajectory planning. Using score-based diffusion models, this work implements a novel generative framework to generate ballistic transfers from Earth to Mars. We further analyze the model's ability to learn the characteristics of the original dataset and its ability to produce transfers that follow the underlying dynamics. Ablation studies were conducted to determine how model performance varies with model size and trajectory temporal resolution. In addition, a performance benchmark is designed to assess the generative model's usefulness for trajectory design, conduct model performance comparisons, and lay the groundwork for evaluating different generative models for trajectory design beyond diffusion. The results of this analysis showcase several useful properties of diffusion models that, when taken together, can enable a future system for generative trajectory design powered by diffusion models.