I am a Ph.D. student at the Automatic Control Laboratory at ETH Zürich under the supervision of Professor John Lygeros and Professor Tobias Sutter. My research focuses on the control of stochastic systems under safety constraints, which includes the analysis of reachability, invariance, and chance-constrained optimal control problems. Therefore, I utilize tools from dynamic programming and convex optimization.
Real-world systems are often subject to uncertainties and disturbances, which may stem from external influences or dynamics which are hard to model. Accordingly, these effects are often difficult to predict and approximated as stochastic perturbations, which can lead to an unsafe control behavior if not properly managed.
I research control approaches for stochastic systems which maximize the probability of safety. My main focus is on the development of scalable algorithms for the analysis and control of high-dimensional systems. Therefore, I developed theory on approximate dynamic programming approaches as well as stochastic control architectures.
Selected publications:
Beyond safety, systems are usually subject to performance objectives. However, there might be trade-offs between safety and performance. Designing controllers which trade-off risks against performance are challenging to compute, as they pose non-Markov constraints.
I proved that such problems can be reformulated as Markov Decision Processes on an augmented state space, which allows solving them using (approximate) dynamic programming approaches.
Selected publications:
Related to my theoretical work, I am interested in practical applications involving biological systems, such as medical instruments and precision agriculture. Not only do such systems have highly nonlinear dynamics, making them challenging to control. But they are also subject to biological variability, introducing uncertainties in the evolution of the system state.
I am currently working on the design and control of a hydroponics system, which is a soil-less cultivation method that allows for a precise control over plant growth conditions. The goal is to develop an optimal, learning-based control strategy that maximizes crop yield vs. resource usage (water, electricity). I further study the optimal fertilization of fields. Publications on these topics will become available in 2026.
All of my code & publications are open-access! Manuscripts: Arxiv, Code: GitHub.
Teaching has a high priority for me. I believe that it is our responsibility to share our fascination and knowledge to inspire future generations of students in the same way that it has been shared with us.
I usually have several open student projects available. Feel free to contact me at any time if you are looking for a Bachelor-, Semester-, or Master-project based around control theory and hands-on applications.
| Name | Title | Type | Year | Co-Supervisor |
|---|---|---|---|---|
| Oliver Baumgartner | Optimal Remembering with Memoryless Policies for Partially Observable Markov Decision Processes | BT | 2026 | Liam Cregg, Kevin Wallington |
| Oliver Baumgartner Samuel Baumgartner Louis Miller David Lammering Florian Borchard |
Interactive Control Research Communication using the Ball-on-Plate system | GP | 2025 | André Warnecke |
| Peter Schmitt | Meta-Learning Model Predictive Controllers for Unseen Tasks and Changing Environments | SP | 2025 | Riccardo Zuliani, Gabriele Fadini |
| Salma Elfeki | High-Performance MPC Controllers Under Environmental Changes | MT | 2025 | Riccardo Zuliani |
| Rye Gleason | Optimal Crop Fertilization Control Strategies and Verification | MT | 2025 | Kevin Wallington |
| Tristan Zeller | Joint Chance Constrained Reinforcement Learning | SP | 2025 | Marta Fochesato |
| Daniel Böhler | Iterative Learning Control for Plants in Controlled Environments | SP | 2025 | Kevin Wallington |
| Jair Reyes | Optimal Control of Plants in Controlled Environments | SP | 2025 | Kevin Wallington |
| Nicola Scagnetti | Design and Control of a Hydroponics System | MT | 2025 | |
| Jonathan Hilberg | Model Free Joint Chance Constrained Optimal Control | MT | 2024 | Marta Fochesato |
| Andrey Kharitenko | Robust Covariance Steering | MT | 2024 | Marta Fochesato, Anastasios Tsiamis |
| Simon Dreher | Redesign of an Hydroponics System | SP | 2024 | |
| Ariel Bergmann | Control of an Inverted Pendulum | BT | 2024 | |
| Alexander Kaspar | Meta-Learning Model Predictive Control for the Ball-On-A-Plate System | SP | 2024 | Jiaqi Yan, Riccardo Zuliani |
| Mirco Vandeventer | Meta-Learning Reinforcement Learning for the Ball-On-A-Plate System | SP | 2024 | Jiaqi Yan |
| Simon Frölich | Hardware Reconfiguration and Control of the Ball-On-A-Plate System | BT | 2024 | |
| Gian-Andrin Coolen | Hydroponics - Mechanics | BT | 2024 | |
| Oscar Kläsi | Hydroponics - Electronics | BT | 2024 | |
| Leo Noth | Hydroponics - Sensing and Control | BT | 2024 | |
| Daniel Böhler* | Hydroponics - Software Design and Control | BT | 2024 | |
| Alexander Kaspar | MPC of the Mounted-Helicopter Experiment | BT | 2023 | |
| Elias Bai | Improved Reinforcement Learning Control of the Ball-On-A-Plate System | BT | 2023 | |
| Joram Ebinger | LQR and Reinforcement Learning Control of a Ball-On-A-Plate System | BT | 2023 | |
| Gentrit Gashi | Water-Level Sensing for the Quad-Tank Experiment | BT | 2023 |
GP: Group Project, BT: Bachelor Thesis, SP: Semester Project, MT: Master Thesis
* Awarded Best Bachelor Thesis 2024 by the Swiss Control Association (Schweizer Gesellschaft für Automatik).
Research means responsibility. We promise that our work has real-world impact, and paves the way for a sustainable, healthy and prosperous future. Therefore, we receive tax money. It is our responsibility to show the public that their investment is well spent. Further, sharing our results and fascination for innovation is a unique opportunity to inspire future generations of students and researchers.
To communicate the foundations and importance of control, I designed the Ball-on-Plate system, a motorized ball-balancing screen. The system displays the basic principles of various control algorithms in interactive games and will be exhibited at several Swiss museums in 2026. Details are available on the Ball-on-Plate website. The following is a short sneak peek of some controllers we implemented for a museum exhibition.
Together with Jared Miller, I wrote a two-part blog post for the NCCR Automation on the topic of safe automation targeted to the general public and undergraduate students. The blog posts explains the importance of safety in automation, and principles by which controllers achieve safety. It is a great read for anyone interested in learning about the safe control and getting a first overview in an accessible way. Check it out here: Part 1 and Part 2
I am grateful for the financial support I have received during my Ph.D. studies, which enabled me to pursue my vision in research and teaching. I was supported by the European Research Council under grant 787845 (OCAL) and the Swiss National Science Foundation under NCCR Automation under grant 51NF40_225155. I further received the following grants and funding.
| Amount | Source | Description | Year |
|---|---|---|---|
| 85,735 CHF | KIM (Commission of Computer Science, ETH Zürich) | Grant for designing a hands-on control class using cart-pole systems | 2024 |
| 20,000 CHF | NCCR Automation, Swiss National Science Foundation | Grant for my research visit at MIT, Boston, USA | 2024 |
| - | ACC (American Control Conference) | Student Travel Grant | 2022 |
