Research Assistant


Foto von Jonas Umlauft

M.Sc. Jonas Umlauft

Technische Universität München

Lehrstuhl für Informationstechnische Regelung (Prof. Hirche)

Postadresse

Postal:
Barerstr. 21
80333 München

Short Biography:

  • 05/2015 - present: PhD candidate at Chair of Information-Oriented Control (ITR), Technical University of Munich (TUM), Germany
  • 08/2015 - present: Doctoral representative electrical engineering and member of the TUM Graduate Council 
  • 10/2013 - 03/2015: M.Sc. in Electrical and Information Engineering, Technical University of Munich (TUM), Germany; Thesis on extensions to PILCO supervised by Carl E Rasmussen
  • 05/2011 - 09/2013: B.Sc. in Electrical and Information Engineering, Technical University of Munich (TUM), Germany;  Thesis on cooperative manipulation supervised by Dominik Sieber

Full CV

Research Interests

  • System identification using machine learning
  • Control based on probabilistic dynamical models
  • Risk-sensitive control using Gaussian Processes

Selected Publications & Public Code

An Uncertainty-based Control Lyapunov Approach for Control-affine Systems Modeled by Gaussian Process

Data-driven approaches in control allow for identification of highly complex dynamical systems with minimal prior knowledge. However, properly incorporating model uncertainty in the design of a stabilizing control law remains challenging. Therefore, this letter proposes a control Lyapunov function framework which semiglobally asymptotically stabilizes a partially unknown fully actuated control affine system with high probability. We propose an uncertainty-based control Lyapunov function which utilizes the model fidelity estimate of a Gaussian process model to drive the system in areas near training data with low uncertainty. We show that this behaviormaximizes the probability that the system is stabilized in the presence of power constraints using equivalence to dynamic programming. A simulation on a nonlinear system is provided. 

The code for the publication "An Uncertainty-based Control Lyapunov Approachfor Control-affine Systems Modeled by Gaussian Process" by Jonas Umlauft, Lukas Pöhler and Sandra Hirche published in IEEE Control Systems Letters (L-CSS) and IEEE Conference on Decision and Control (CDC) in 2018 is available here.

Feedback Linearization based on Gaussian Processes with event-triggered Online Learning

Combining control engineering with nonparametric modeling techniques from machine learning allows to control systems without analytic description using data-driven models. Most existing approaches separate learning, i.e. the system identification based on a fixed dataset, and control, i.e. the execution of the model-based control law. This separation makes the performance highly sensitive to the initial selection of training data and possibly requires very large datasets. This article proposes a learning feedback linearizing control law using online closed-loop identification. The employed Gaussian process model updates its training data only if the model uncertainty becomes too large. This event-triggered online learning ensures high data efficiency and thereby reduces the computational complexity, which is a major barrier for using Gaussian processes under real-time constraints. We propose safe forgetting strategies of data points to adhere to budget constraint and to further increase data-efficiency. We show asymptotic stability for the tracking error under the proposed event-triggering law and illustrate the effective identification and control in simulation.

The code for the publication "Feedback Linearization based on Gaussian Processes with event-triggered Online Learning" by Jonas Umlauft and Sandra Hirche accepted to the IEEE Transactions on Automatic Control( TAC) in 2019 is available here.

Video: Smart Forgetting for Safe Online Learning with Gaussian Processes

Learning Stable Gaussian Process State Space Models

Data-driven nonparametric models gain importance as control systems are increasingly applied in domains where classical system identification is difficult, e.g., because of the system’s complexity, sparse training data or its probabilistic nature. Gaussian process state space models (GP-SSM) are a data-driven approach which requires only high-level prior knowledge like smoothness characteristics. Prior known properties like stability are also often available but rarely exploited during modeling. The enforcement of stability using control Lyapunov functions allows to incorporate this prior knowledge, but requires a data-driven Lyapunov function search. Therefore, we propose the use of Sum of Squares to enforce convergence of GP-SSMs and compare the performance to other approaches on a real-world handwriting motion dataset.

The code for the publication "Learning Stable Gaussian Process State Space Models" by Jonas Umlauft, Armin Lederer and Sandra Hirche published at the IEEE American Control Conference (ACC) 2018 is available here.

Uniform Error Bounds for Gaussian Process Regression with Application to Safe Control

Data-driven models are subject to model errors due to limited and noisy training data. Key to the application of such models in safety-critical domains is the quantification of their model error. Gaussian processes provide such a measure and uniform error bounds have been derived, which allow safe control based on these models. However, existing error bounds require restrictive assumptions. In this paper, we employ the Gaussian process distribution and continuity arguments to derive a novel uniform error bound under weaker assumptions. Furthermore, we demonstrate how this distribution can be used to derive probabilistic Lipschitz constants and analyze the asymptotic behavior of our bound. Finally, we derive safety conditions for the control of unknown dynamical systems based on Gaussian process models and evaluate them in simulations of a robotic manipulator.

The code for the publication "Uniform Error Bounds for Gaussian Process Regression with Application to Safe Control" by Armin Lederer, Jonas Umlauft and Sandra Hirche accepted to the Conference on Neural Information Processing Systems (NeurIPS) is available here.

Additional Material

Video: Simulation of Stable Gaussian Process based Tracking Control

Video: Robot manipulator with Gaussian Process based Tracking Control

Past Student Projects

  • BA: Risk-Sensitive Cooperative Dynamic Movement Primitives using Gaussian Processes [PDF]
  • BA: Human-Human Cooperation Behaviour Analysis using Uncertainty Models [PDF]
  • BA: Enhancing Variance-Dependent Cooperative Dynamic Movement Primitives [PDF]
  • BA: Learning Stochastic Stable Systems [PDF]
  • HS: Trajectory Tracking of learned Dynamics [PDF]
  • HS: Gaussian Processes for Model Predictive Control [PDF]
  • BA: Online Kinematic Teaching using Optimal Control and Gaussian Processes [PDF]
  • IP: Learning stable human goal directed movements [PDF]
  • HS: Learning on Manifolds for Stable Dynamical System [PDF]
  • FP: Modelling Uncertainties using Wishart Processes [PDF]
  • PP: Learning Stochastic Stable Systems using Sum of Squares Control Lyapunov Functions [PDF]
  • BA: Indirect Adaptive Control based on Gaussian Processes [PDF]
  • BA: Computed Torque Control with Gaussian Process Regression for Robotics [PDF]
  • BA: Learning and Control for Stochastic Stable Systems [PDF]
  • FP: Comparing Multi-Step Predictions for Gaussian Processes [PDF]
  • MA: Data-Driven Approaches to Model Predictive Control [PDF]
  • BA: Path Integral Control for Gaussian Processes [PDF]
  • BA: Efficient Exploration for Gaussian Process Models [PDF]
  • FP: Enhancing Uncertainty-based Control for Gaussian Processes [PDF]
  • BA: Learning Control for Gaussian Process Models [PDF]
  • FP: Efficient Exploration of Dynamical Systems [PDF]
  • BA: Learning Control for Robotic Manipulators [PDF]
  • MA: Identification and Control for Input-Output Systems based on Gaussian Processes [PDF]

Projects

ERC Starting Grant: Control based on Human Models

Publications

2020

  • A. Capone; G.Noske; J. Umlauft; T. Beckers; A. Lederer; S. Hirche: Localized active learning of Gaussian process state space models. Learning for Dynamics & Control, 2020 mehr… BibTeX
  • A. Capone; A. Lederer; J. Umlauft; S. Hirche: Data Selection for Multi-Task Learning Under Dynamic Constraints. IEEE Control Systems Letters 5 (3), 2020, 959-964 mehr… BibTeX
  • A. Capone; G. Noske; J. Umlauft; T. Beckers; A. Lederer; S. Hirche: Localized Active Learning of Gaussian Process State Space Models. 2020 mehr… BibTeX
  • A. Lederer; A. Capone; J. Umlauft; S. Hirche: How Training Data Impacts Performance in Learning-based Control. IEEE Control Systems Letters, 2020, 1-1 mehr… BibTeX
  • J. Umlauft; S. Hirche: Learning Stochastically Stable Gaussian Process State-Space Models. IFAC Journal of Systems and Control 12, 2020 mehr… BibTeX
  • J. Umlauft; S. Hirche: Feedback Linearization based on Gaussian Processes with event-triggered Online Learning. IEEE Transactions on Automatic Control, 2020 mehr… BibTeX
  • J. Umlauft; T. Beckers; A. Capone; A. Lederer; S. Hirche: Smart Forgetting for Safe Online Learning with Gaussian Processes. Learning for Dynamics & Control, 2020 mehr… BibTeX

2019

  • A. Lederer; J. Umlauft; S. Hirche: Posterior Variance Analysis of Gaussian Processes with Application to Average Learning Curves. arXiv preprint: arXiv:1906.01404, 2019 mehr… BibTeX
  • A. Lederer; J. Umlauft; S. Hirche: Uniform Error Bounds for Gaussian Process Regression with Application to Safe Control. Conference on Neural Information Processing Systems (NeurIPS), 2019 mehr… BibTeX
  • A. Lederer; J. Umlauft; S. Hirche: Uniform Error Bounds for Gaussian Process Regression with Application to Safe Control. Conference on Neural Information Processing Systems (NeurIPS), 2019 mehr… BibTeX

2018

  • J. Umlauft; L. Pöhler; S. Hirche: An Uncertainty-Based Control Lyapunov Approach for Control-Affine Systems Modeled by Gaussian Process. IEEE Control Systems Letters 2 (3), 2018, 483-488 mehr… BibTeX
  • J. Umlauft; T. Beckers; S. Hirche: A Scenario-based Optimal Control Approach for Gaussian Process State Space Models. Proceedings of the European Control Conference (ECC), 2018 mehr… BibTeX
  • L. Pöhler; J. Umlauft; S.Hirche: Uncertainty-based Human Motion Tracking with Stable Gaussian Process State Space Models. IFAC Conference on Cyber-Physical & Human Systems (CPHS), 2018 mehr… BibTeX
  • T. Beckers; J. Umlauft; S. Hirche: Mean Square Prediction Error of Misspecified Gaussian Process Models. Proceedings of the 57th Conference on Decision and Control (CDC), 2018 mehr… BibTeX

2017

  • J. Umlauft; A. Lederer; S. Hirche: Learning Stable Gaussian Process State Space Models. American Control Conference (ACC), IEEE, 2017 mehr… BibTeX
  • J. Umlauft; S. Hirche: Learning Stable Stochastic Nonlinear Dynamical Systems. International Conference on Machine Learning (ICML), 2017 mehr… BibTeX
  • J. Umlauft; T. Beckers; M. Kimmel; S. Hirche: Feedback Linearization using Gaussian Processes. Proceedings of the Conference on Decision and Control (CDC), IEEE, 2017 mehr… BibTeX
  • J. Umlauft; Y. Fanger; S. Hirche: Bayesian Uncertainty Modeling for Programming by Demonstration. International Conference on Robotics and Automation (ICRA), IEEE, 2017 mehr… BibTeX
  • T. Beckers; J. Umlauft; D. Kulić; S. Hirche: Stable Gaussian Process based Tracking Control of Lagrangian Systems. Proceedings of the 56th Conference on Decision and Control (CDC), 2017 mehr… BibTeX
  • T. Beckers; J. Umlauft; S. Hirche: Stable Model-based Control with Gaussian Process Regression for Robot Manipulators. Proceedings of the 20th IFAC World Congress, 2017 mehr… BibTeX

2016

  • Y. Fanger; J. Umlauft; S. Hirche: Gaussian Processes for Dynamic Movement Primitives with Application in Knowledge-based Cooperation. International Conference on Intelligent Robots and Systems (IROS), 2016 mehr… BibTeX

2014

  • J. Umlauft; D. Sieber; S. Hirche: Dynamic Movement Primitives for Cooperative Manipulation and Synchronized Motions. IEEE International Conference on Robotics and Automation (ICRA), 2014, 6 mehr… BibTeX