Guanru Pan, postdoctoral researcher at the Institute of Control Systems at Hamburg University of Technology, and his co-authors present a data-driven framework for uncertainty propagation under unmeasured or statistically unmodeled disturbances. By introducing residual disturbances, which consolidate all unstructured effects into a single quantity that can be estimated from data, the authors obtain a stochastic predictor that is both causal and distributionally consistent under mild assumptions. The framework further enables efficient uncertainty quantification through polynomial chaos expansions and higher-order Chebyshev inequalities.
The proposed approach was validated using experimental data from a smart home in Norway. The study was carried out in collaboration with Dirk Reinhardt and Sebastien Gros from the Department of Engineering Cybernetics at the Norwegian University of Science and Technology (NTNU) in Trondheim.
Guanru Pan, Dirk Reinhardt, Sebastien Gros, Timm Faulwasser (2026). Uncertainty Propagation under Residual Disturbances: A Smart-Home Case Study. Accepted for the IFAC World Congress 2026, Busan, Korea. Preprint: arXiv:2605.15851.
