Authors: Giacomo Acciarini, Atılım Güneş Baydin, Dario Izzo

Published on: February 07, 2024

Impact Score: 8.15

Arxiv code: Arxiv:2402.04830

Summary

• What is new: Introduction of dSGP4, a differentiable version of the SGP4 orbital propagation method, implemented in PyTorch for improved precision and integration with neural networks.
• Why this is important: SGP models, while fast and reliable for predicting the positions and velocities of Earth-orbiting objects, lack the precision compared to numerical propagators.
• What the research proposes: By making SGP4 differentiable and integrating it with neural networks, dSGP4 improves precision and supports advanced space-related applications.
• Results: dSGP4 achieves higher precision in orbital prediction by allowing fine-tuning with ephemeris data, leveraging computational power across CPUs and GPUs for parallel processing.

Technical Details

Technological frameworks used: PyTorch

Models used: Neural networks integrated with the orbital propagator

Data used: Two-Line Element Sets (TLEs), ephemeris data

Potential Impact

Satellite operators, aerospace industry, space research organizations

Want to implement this idea in a business?

We have generated a startup concept here: OrbitAI.