- Learning-Based Optimal Control with Performance Guarantees for Unknown Systems with Latent States. 2024 European Control Conference (ECC), 2024, 90-97 more… BibTeX Full text ( DOI ) Full text (mediaTUM)
M.Sc. Robert Lefringhausen
- Phone: +49 (89) 289 - 25738
- Room: 0305.04.508
- robert.lefringhausen@tum.de
Short Biography
Since 03/2022 | Research Assistant Chair of Information-oriented Control Technical University of Munich, Germany |
04/2019 – 01/2022 | Master of Science, Electrical Engineering and Information Technology Technical University of Munich, Germany Focus: control, machine learning & robotics |
09/2019 – 01/2020 | Semester abroad Linköping University, Sweden |
10/2015 – 02/2019 | Bachelor of Science, Electrical Engineering and Information Technology Karlsruhe Institute of Technology, Germany |
Research Interests
- Learning-Based Control – Integrating methods from machine learning and control theory to develop data-driven controllers for complex dynamical systems.
- Uncertainty Quantification and Robustness – Developing techniques to reason about uncertainty and enable safe and reliable decision-making, with a focus on Bayesian approaches and Monte Carlo methods.
- Learning-Based Control of Unknown Systems with Latent States – Designing algorithms for joint state and dynamics estimation, as well as optimal control, in the absence of full state measurements using Bayesian state-space models and (particle) Markov chain Monte Carlo methods to infer analytically intractable distributions.
- Active Learning for Dynamical Systems – Creating strategies that efficiently explore and identify system dynamics from limited data, guided by model uncertainty.
Research Project
I am currently involved in the ONE MUNICH Strategy Forum Project - Next generation Human-Centered Robotics.
Student Projects and Theses
I am constantly looking for motivated students who are interested in my field of research. Please contact me via e-mail if you are interested in working on a thesis (e.g., bachelor or master thesis) under my supervision, even if no open topics are currently listed.
Please include your preferred starting date as well as your CV and transcript of records in your e-mail. This helps me to select a topic matching your background.
Currently available theses
- Fusion of Model Predictive Control and Reinforcement Learning for the Safe Exploration of Unknown Systems with Latent States
[PDF]
Ongoing theses
- Xinyi Shao: Learning Bayesian State-Space Models with Latent States via Likelihood Optimization (master thesis)
- Supitsana Srithasan: Safe Exploration in Partially Observable Systems via Model Predictive Control (research internship)
Past Theses
- Radu-Andrei Bourceanu: Neural Network Policies for the Approximate Model Predictive Control of Unknown Systems with Latent States (bachelor thesis)
- Lukas Hochschwarzer: Kernel Embedding for Particle Gibbs-Based Optimal Control (master thesis)
- Aymen Nasri: Kernel Embedding for Particle Gibbs-Based Optimal Control (research internship)
- Sami Noel: Stability Certificates for Unknown Systems with Latent States (research internship)
- Raphael Nonnenmann: Evaluation of Neural Network-Based Approximations for Real-Time Model Predictive Control of Unknown Systems with Latent States (bachelor thesis)
- Joshua Schenk: Fusion of Model Predictive Control and Reinforcement Learning for the Safe Exploration of Unknown Systems with Latent States (research internship)
- Supitsana Srithasan: Utilizing Particle Gibbs Samples for Scenario-Based Model Predictive Control (bachelor thesis)
- Sarah Weber: Particle Markov Chain Monte Carlo-Based Optimal Control for Vertical Farming (bachelor thesis)
- Keyu Xuan: Particle Gibbs Sampling Based on Multiple Trajectories (engineering practice)
- Yi Yue: Towards Online Learning and Control of Unknown Systems with Latent States (bachelor thesis)
- Kevin Zierenberg: Parallel Implementation of a Particle Gibbs Based Control Algorithm on a Graphics Processing Unit (bachelor thesis)
- Zhiqian Zhou: Active Learning of Unknown Dynamics with Latent States under Constraints (bachelor thesis)
- Zhiqian Zhou: Hyperparameter Tuning for Gaussian Process State-Space Models with Latent States (engineering practice)