Design an early warning system showing how orbital factors and space weather influence the satellite drag using a Machine Learning (ML) model
Satellites in Low Earth Orbit (LEO) face a persistent threat: atmospheric drag driven by space weather causes progressive orbital decay, thereby shortening satellite lifetimes and creating costly operational risks.
Build an ML-powered early warning system that predicts satellite decay rates using orbital parameters and geomagnetic activity data.
Six years of Two-Line Element (TLE) data of LEO satellites operated by Planet, from Space-Track were combined with geomagnetic and solar activity indices obtained from GFZ Helmholtz Center for Geosciences, to model how space weather impacts satellite lifetimes.
| Achievement | Detail |
|---|---|
| Prediction Accuracy | 0.90–0.92 for select orbital groups |
| Decay Trend Modeling | Semi-major axis trends tracked across multiple years |
| Feature Importance | Solar flux and geomagnetic storms identified as top drivers |
| Interactive Tool | Decay Rate Predictor — input satellite parameters, get risk output |
- Maneuver satellites proactively to extend operational lifetimes
- Increase space weather monitoring across mission planning workflows
- Develop time-series models for higher-resolution future decay forecasting
Predictive modeling can protect satellite operators, scientists, and commercial users from costly orbital risks, while supporting the long-term sustainability of Low Earth Orbit.