About the project
This project will use machine learning and big data techniques to develop real-time satellite orbit predictions for safer, smarter space traffic management. It aims to integrate high-precision laser ranging satellite tracking data to continuously update the biggest unknown, atmospheric density. This can improve tracking accuracy by up to an order of magnitude.
The rise of satellite mega-constellations and improved tracking systems has led to an explosion of precise satellite data. This project offers the chance to help solve one of the biggest challenges in modern space operations: managing an increasingly congested low Earth orbit (LEO) environment.
Working with leading academics and our industry partner Lumi Space, you will develop a world-class, high-precision orbit prediction model. The goal is to reduce false collision alerts and enable satellite operators to act decisively on real threats. This is critical for safe and efficient space traffic management.
At the heart of the project is a responsive, high-fidelity atmospheric density model, continuously updated using real-time tracking data. You’ll address two main challenges:
- model design: identifying and training a compact, accurate representation of atmospheric density using reduced-order models, time series methods, and machine learning approaches such as physics-informed neural networks
- real-time data assimilation: using sequential filtering techniques, such as high-order Kalman filters, to fuse live tracking data into model-predictive control for orbit forecasting
Throughout, you’ll work closely with Lumi Space, gaining valuable industrial experience and access to real satellite tracking data to validate your work.
In addition to the supervisory team at the University of Southampton, you will be supported by the following external supervisor:
- Dr Peter Bartram (Lumi Space)
