Embedded Files
Research
Research
Modern applications such as autonomous driving, soft robotics, and human–robot interaction push today’s technical systems to unprecedented levels of complexity. As a result, traditional modeling and control approaches are often prohibitively time-consuming or even infeasible. Data-driven methods offer a powerful alternative: they require little expert tuning and have demonstrated impressive performance. Yet their lack of guarantees has so far limited their use in safety-critical systems.
Our research addresses this gap by advancing safe, learning-based modeling and control of physical systems. We develop novel data-driven control methods that combine high performance with rigorous safety guarantees in closed-loop operation. A key focus is on integrating physical prior knowledge into learning algorithms, improving their generalizability, reliability, and interpretability. Ultimately, our work contributes to the next generation of safe, robust, and intelligent control technologies for complex physical systems.
News
News
12/2025 CDC in Rio de Janeiro was a great success! We organized a workshop and an invited session on physics-informed learning for control and presented our paper on passivity-based trajectory tracking control.
11/2025 We will organize the DESTIN Workshop and a tutorial session on physics-informed learning at the next CPS-IoT Week. Stay tuned for more!
09/2025 Our article "A Safety-Driven Interpretable Model for Vehicle Control with Impact on Traffic" has been accepted for publications at Transactions on Intelligent Transportation Systems.
09/2025 I will be serving as a Guest Associated Editor for the OJ-CSYS special section on Intersection of Machine Learning with Control
08/2025 We’ll be more involved than ever in physics-informed ML at this year’s CDC, with a workshop, an invited session, and a new paper on PIML for control.
07/2025 Kicking off the ACC Workshop on Physics-informed learning for control in Denver.
05/2025 I will present our recent work on Physics-informed system identification at the Workshop on Nonlinear System Identification Benchmarks
04/2025 Check out our new tutorial paper on "Safe Physics-Informed Machine Learning for Dynamics and Control" (will be presented at ACC 2025)
04/2025 Congratulations to Peilun Li and Kaiyuan Tan on the acceptance of their paper "NAPI-MPC: Neural Accelerated Physics-Informed MPC for Nonlinear PDE Systems" at L4DC 2025.
01/2025 Congratulations to Kaiyuan Tan and Peilun Li on the acceptance of their paper "PnP-PIML: Physics-informed Learning of Outlier Dynamics using Uncertainty Quantified Port-Hamiltonian Models" at ICRA 2025.
12/2024 Congratulations to Peilun Li who was selected for Honorable Mention for the 2024-2025 CRA Outstanding Undergraduate Researcher Award (URA).
12/2024 I will present at the CDC 2024 workshop Learning Dynamics from Data: Fusing Machine Learning and System Identification
10/2024 Tobias Wolff from University of Hannover is visiting our group. Welcome!
08/2024 Our paper Interpretable Finite State Machine Controller: A Case Study on Lane Merge Yield Mode has been accepted at ITSC 2024.
08/2024 I presented at the Port-Hamiltonian Systems Seminar series at the University of Wuppertal, Germany.
07/2024 Our group has three presentations at the 16th World Congress on Computational Mechanics in Vancouver.
07/2024 We're welcoming our new team members Mohammad Ali and Kendra Givens!
07/2024 We presented our work on Physics-informed learning of PDEs with uncertainty quantification at the L4DC in Oxford.
07/2024 We organized a workshop on physics-informed learning with fantastic talks and a great tutorial session at the American Control Conference (ACC).
06/2024 Keeping strong ties with the European control community by presenting our recent work at the ECC Workshop on Physics-Informed Learning in Control!
06/2024 We presented our new Python toolbox for Bayesian learning of Port-Hamiltonian systems at the 8th IFAC Workshop on Lagrangian and Hamiltonian Methods for Non Linear Control
05/2024 I gave a presentation at the Artificial Intelligence for Robust Engineering & Science Workshop (AIRES), PNNL
04/2024 I gave a virtual presentation at the UNC Charlotte Robotics and Controls Seminar series
04/2024 Peilun Li and Kaiyuan Tan got a paper accepted at the 8th IFAC Workshop on Lagrangian and Hamiltonian Methods for Nonlinear Control.
02/2024 Peilun Li and Kaiyuan Tan will present their most recent works on physics-informed learning at the 16th World Congress on Computational Mechanics
01/2024 We will organize a workshop on Physics-informed Machine Learning for Modeling, Control, and Optimization at the American Control Conference, July 10-12, 2024.
Google Sites
Report abuse