The First Open-Source Framework for Learning Stability Certificates from Data
Abstract
Before 2025, no open-source system existed that could learn Lyapunov stability certificates directly from noisy, real-world flight data. This work addresses that gap by proposing a data-driven approach that learns Lyapunov functions from trajectory data under realistic, noise-corrupted conditions. Unlike statistical anomaly detectors that only flag deviations, the proposed method assesses whether the system can still be certified as stable. Applied to public data from the 2024 SAS severe turbulence incident, this framework revealed that, within 60 seconds of the aircraft's descent becoming abnormal, no Lyapunov function could be constructed to certify system stability. To the best of our knowledge, this is also the first application of a data-driven Lyapunov-based stability verification method to real civil aviation data, achieved without any access to proprietary controller logic. The proposed framework is open-sourced and available at: https://github.com/HansOersted/stability
Cite
@article{arxiv.2509.20392,
title = {The First Open-Source Framework for Learning Stability Certificates from Data},
author = {Zhe Shen},
journal= {arXiv preprint arXiv:2509.20392},
year = {2025}
}
Comments
Accepted by IEEE Aerospace Conference