In many practical control applications, the mathematical model of the system to be controlled or the environment of the system is unknown or uncertain, making the use of most standard control designs impossible. Data-based control is the area of research that investigates the development of stabilizing controllers using only suitable collected input-output information, without any knowledge of the system model. Similarly, learning-based control uses artificial intelligence or online learning algorithms to learn the model or the environment of the system or directly determine a stabilizing controller from input-output data. These control strategies have strong potential applications for control of nonlinear systems, where obtaining accurate system models is often difficult or practically unfeasible.
In particular, our current research focuses on
- online convex optimization for control of dynamical systems,
- data-based descriptions of nonlinear systems,
- the development of data-driven predictive controllers,
- the design of off-policy reinforcement learning algorithms to obtain state-feedback controllers,
- the study of data-based controllers for irregularly sampled systems, and
- the development of data-/learning-based estimators.
Selected Publications
-
(2023): Efficient Off-Policy Q-Learning for Data-Based Discrete-Time LQR Problems, IEEE Transactions on Automatic Control, pp. 1-12
DOI: 10.1109/TAC.2023.3235967 -
(2023): On the design of persistently exciting inputs for data-driven control of linear and nonlinear systems, IEEE Control Systems Letters, vol. 7, pp. 2629 - 2634
DOI: 10.1109/LCSYS.2023.3287133
arXiv: 2303.08707 -
(2023): Data-Based Control of Feedback Linearizable Systems, IEEE Transactions on Automatic Control (Early Access)
DOI: 10.1109/TAC.2023.3249289
arXiv: 2204.01148 -
(2023): Data-driven nonlinear predictive control for feedback linearizable systems, 22nd IFAC World Congress: Yokohama, Japan, July 9-14, 2023. IFAC-PapersOnLine, Vol. 56, Issue 2, pp. 617-624
DOI: 10.1016/j.ifacol.2023.10.1636
arXiv: 2211.06339 -
(2023): On the relation between dynamic regret and closed-loop stability, Systems & Control Letters, vol. 177, p. 105532
DOI: 10.1016/j.sysconle.2023.105532
arXiv: 2209.05964 -
(2023): An efficient off-policy reinforcement learning algorithm for the continuous-time LQR problem, Accepted for IEEE 62nd Conference on Decision and Control (CDC)
arXiv: 2303.17819 -
(2022): Data-Based Moving Horizon Estimation for Linear Discrete-Time Systems, 2022 European Control Conference (ECC), pp. 1778-1783
DOI: 10.23919/ecc55457.2022.9838331
arXiv: 2111.04979 -
(2022): On a continuous-time version of Willems' lemma, IEEE 61st Conference on Decision and Control (CDC), pp. 2759-2764
DOI: 10.1109/CDC51059.2022.9992347 -
(2022): Online Convex Optimization for Data-Driven Control of Dynamical Systems, IEEE Open Journal of Control Systems, vol. 1, pp. 180-193
DOI: 10.1109/OJCSYS.2022.3200021 -
(2021): Data-Driven Model Predictive Control With Stability and Robustness Guarantees, IEEE Transactions on Automatic Control, 2021, Vol. 66, No. 4, pp. 1702-1717
DOI: 10.1109/TAC.2020.3000182
Selected Projects
-
Cont4Med - Estimation and control under limited information with application to biomedical systemsLed by: Prof. Dr.-Ing. Matthias MüllerTeam:Year: 2021Funding: This project has received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (grant agreement No 948679).Duration: 2021 - 2025
-
Convex online optimization for dynamic systems controlLed by: Prof. Dr.-Ing. Matthias MüllerTeam:Year: 2023Funding: Deutsche Forschungsgemeinschaft (DFG) - 505182457Duration: 2023 - 2025