Skip to content
Technical University of Munich
Picture of Nicolas Hoischen

ETH M.Sc. Nicolas Hoischen

Technical University of Munich

Chair of Information-oriented Control (Prof. Hirche)

Google Scholar
LinkedIn

Short Biography

  • 09/2023 - present: PhD candidate at Chair of Information-Oriented Control (ITR),
    Technical University of Munich (TUM), Germany.
     
  • 06/2023: M.Sc. in Information Technology and Electrical Engineering, 
    Eidgenössische Technische Hochschule Zürich (ETH Zürich), Switzerland.
    • Thesis: Safe Connectivity Maintenance for a Fleet of Underactuated Seaweed Farms in Dynamic Oceanic Environments,
      UC Berkeley, USA.
       
  • 08/2020: B.Sc. in Electrical and Electronic Engineering, 
    École Polytechnique Fédérale de Lausanne (EPFL), Switzerland.   
    • Thesis: Modelling, Simulation and Control of Eirsat-1 in Orbit,
      University College Dublin (UCD), Ireland.

Research

I am involved in the Seaclear2.0 Project and affiliated with the relAI School.

Learning Optimal Control

Learning the swing-up task for the Cartpole system (Master Thesis: Tobias Wittmann)

 

My research focuses on enabling autonomous decision-making in cyber-physical systems to optimize their performance and behavior. 
 
I am particularly interested in:

  • Modeling and System Identification via Linear Transfer Operators
  • Using Operator-Theoretic Approaches to transform non-convex control problems into convex formulations
  • Data-Driven Control using Kernel Methods
  • Risk Measures for decision-making under uncertainty

Seaclear2.0

First sea trials in Marseille (December 2024). The USV seen in this video is the Seacat2.0 from Subsea Tech (https://www.subsea-tech.com/)

 

In the Seaclear2.0 project, we are developing an autonomous robotic system for seafloor litter collection. Building on Seaclear1.0, we aim to scale up to handle larger and heavier debris using a smart underwater grapple. 
 

My research focuses on controlling the grapple, including developing a propulsion system with position and attitude control for fully autonomous navigation. Additionally, I am working on hardware development to optimize the grasping mechanism for efficient debris retrieval.
 

Contact

I am always looking for motivated students interested in my research. 
If you are considering a thesis in this area, please feel free to email me. Let me know your planned start date, and attach your CV and grade transcript so we can find the right topic for you.

Completed Projects

  • Sami Leon Noel Aziz Hanna: Practical Stabilization of Bilinear Koopman Models via Control Lyapunov Functions (Master Thesis)
  • Xiao, Xuehua: Near-Globally Optimal Approximation-Based Model Predictive Control via Diffusion Models (Master Thesis)
  • Tobias Wittmann: Data-Driven Optimal Feedback Laws for Robotic Systems via Kernelized Generator Learning (Master Thesis)
  • Matthias Laumeyer: Designing and Testing a ROS2 Interface for Thrust Allocation to T500 BlueRobotics Thrusters (Internship)
  • Alisa Pevec: Risk-averse Trajectory Optimization for an ROV with Unknown Payload Mass (Bachelor Thesis)
  • Matthias Laumeyer: Power-Constrained Thrust Allocation for an Overactuated ROV Grapple (Bachelor Thesis)
  • Yanbo Wang: Migration of a SensoJoint Drive ROS1 Interface and Functionalities to ROS2 (Internship)
  • Huang Ziyang: Control of a SensJoint Drive using ROS2 via CANopen over Ethercat (Internship)
     

Selected Publications (Full list on Google Scholar)

2025

  • Nicolas Hoischen, Petar Bevanda, Max Beier, Stefan Georg Sosnowski, Boris Houska, Sandra Hirche: Operator Models for Continuous-Time Offline Reinforcement Learning. TUM Lehrstuhl für Informationstechnische Regelung (Prof. Hirche), 2025, more… BibTeX
  • Origane, Yuki; Hoischen, Nicolas; Huang, Tzu-Yuan; Kurabayashi, Daisuke; Sosnowski, Stefan; Hirche, Sandra: Risk-Aware Trajectory Optimization and Control for an Underwater Suspended Robotic System. , 2025 more… BibTeX
  • Petar Bevanda, Nicolas Hoischen, Tobias Wittmann, Jan Brüdigam, Sandra Hirche and Boris Houska: Kernel-Based Optimal Control: An Infinitesimal Generator Approach. 7th Annual Learning for Dynamics & Control Conference (L4DC 2025), 2025 more… BibTeX
  • Tzu-Yuan Huang, Armin Lederer, Nicolas Hoischen, Jan Brudigam, Xuehua Xiao, Stefan Sosnowski, Sandra Hirche: Toward Near-Globally Optimal Nonlinear Model Predictive Control via Diffusion Models. Proceedings of the 7th Annual Learning for Dynamics & Control Conference, University of Michigan, 2025, 777-790 more… BibTeX

2024

  • Nicolas Hoischen, Petar Bevanda, Stefan Sosnowski, Sandra Hirche, Boris Houska: Data-Driven Stochastic Optimal Control in Reproducing Kernel Hilbert Spaces. Chair of Information-oriented Control, TU München, 2024, more… BibTeX
To top