Validation of Satellite Lifetime Predictions at Leonid Space
Scott Shambaugh

TL;DR
This paper validates satellite lifetime prediction methods using extensive backtesting with historical data, demonstrating significant accuracy improvements over existing tools and establishing a performance baseline for operational use.
Contribution
It introduces a comprehensive validation approach for satellite lifetime prediction, combining probabilistic orbit propagation with space weather forecasts, and demonstrates substantial accuracy and speed improvements.
Findings
Achieved median 1-year prediction accuracy of 6 days under perfect knowledge.
Demonstrated 4x accuracy improvement over ESA's standard tools.
Provided a fast semianalytic propagator enabling large-scale Monte Carlo analysis.
Abstract
We validate Leonid Space's satellite lifetime prediction pipeline through comprehensive backtesting against 934 non-maneuvering satellites that deorbited from LEO between 1961 and 2024. This represents the first large-scale validation of lifetime prediction tooling using forecasted space weather conditions rather than historical hindsight. Our toolchain combines ballistic coefficient estimation from on-orbit data with probabilistic orbit propagation under varying environmental conditions. Using TLE data and space weather records spanning six solar cycles, our three-stage validation approach progressively removes hindsight bias to arrive at fully predictive operational conditions. We achieve 1-year prediction accuracy (median continuously ranked probability score) of 6.0 days (1.6%) under perfect knowledge conditions, 18.6 days (5.1%) with estimated ballistic coefficients and known space…
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
