Department of Electrical Engineering and Automation

Cyber-physical Systems

Cyber-physical systems tightly integrate physical processes with computing and communication. This tight integration enables emerging applications, e.g., coordinating autonomous vehicles or fleets of drones or controlling factory automation machinery over large networks. However, realizing such applications requires developing novel machine learning and control methods. Major challenges stem from (i) the adoption of wireless technology, (ii) the computational limits of embedded devices, and (iii) the unpredictability of the real world.
Cyber-physical Systems
Cyber-physical systems tightly integrate physical processes with computing and communication.

While wireless communication offers unprecedented flexibility in sharing data between systems, which increases collective information and allows collaborative action, it is, in comparison to wired communication, less reliable, and its bandwidth is limited. For example, if all autonomous vehicles in a big city use the same wireless network and communicate simultaneously, the whole network may break down, impeding communication.

Further, for many application examples of cyber-physical systems, such as drones, we need to do computations on lightweight devices, which limits their computational power. This is particularly challenging for machine learning algorithms, which are often demanding in terms of computations. Still, machine learning algorithms are an essential asset of cyber-physical systems. They are essential, especially because cyber-physical systems are supposed to act autonomously in the real world, and not all situations they may encounter can be anticipated at design time. Machine learning methods can bring the required flexibility and adaptability for systems to work safely in unseen situations.

The cyber-physical systems group addresses these challenges by co-designing control, machine learning, and communication, both through theoretical advances and practical experiments.

The cyber-physical systems group is led by assistant professor Dominik Baumann.

Latest publications

A computationally lightweight safe learning algorithm

Dominik Baumann, Krzysztof Kowalczyk, Koen Tiels, Paweł Wachel 2024 2023 62nd IEEE Conference on Decision and Control, CDC 2023

Safe reinforcement learning in uncertain contexts

Dominik Baumann, Thomas B. Schon 2024 IEEE Transactions on Robotics

On the trade-off between event-based and periodic state estimation under bandwidth constraints

Dominik Baumann, Thomas B. Schön 2023 Proceedings of the IFAC World Congress

Reply to “The Limitations of Growth-Optimal Approaches to Decision Making Under Uncertainty”

Oliver Hulme, Arne Vanhoyweghen, Colm Connaughton, Ole Peters, Simon Steinkamp, Alexander Adamou, Dominik Baumann, Vincent Ginis, Bert Verbruggen, James Price, Benjamin Skjold 2023 Econ Journal Watch

Toward Multi-Agent Reinforcement Learning for Distributed Event-Triggered Control

Lukas Kesper, Sebastian Trimpe, Dominik Baumann 2023 Proceedings of the Learning for Dynamics and Control Conference

GoSafeOpt: Scalable Safe Exploration for Global Optimization of Dynamical Systems

Bhavya Sukhija, Matteo Turchetta, David Lindner, Andreas Krause, Sebastian Trimpe, Dominik Baumann 2023 Artificial Intelligence

Automatic nonlinear MPC approximation with closed-loop guarantees

Abdullah Tokmak, Christian Fiedler, Melanie N. Zeilinger, Sebastian Trimpe, Johannes Köhler 2023 arXiv.org

Approximate MPC with Kernel-based Methods

Abdullah Tokmak, Christian Fiedler, Sebastian Trimpe, Melanie N. Zeilinger, Johannes Köhler 2022
More information on our research in the Aalto research portal.
Research portal
  • Published:
  • Updated: