Koopman NMPC: Koopman-based Learning and Nonlinear Model Predictive Control of Control-affine Systems

  title={Koopman NMPC: Koopman-based Learning and Nonlinear Model Predictive Control of Control-affine Systems},
  author={Carl Folkestad and Joel W. Burdick},
  journal={2021 IEEE International Conference on Robotics and Automation (ICRA)},
  • C. Folkestad, J. Burdick
  • Published 17 May 2021
  • Engineering
  • 2021 IEEE International Conference on Robotics and Automation (ICRA)
Koopman-based learning methods can potentially be practical and powerful tools for dynamical robotic systems. However, common methods to construct Koopman representations seek to learn lifted linear models that cannot capture nonlinear actuation effects inherent in many robotic systems. This paper presents a learning and control methodology that is a first step towards overcoming this limitation. Using the Koopman canonical transform, control-affine dynamics can be expressed by a lifted… 

Figures and Tables from this paper

Quadrotor Trajectory Tracking with Learned Dynamics: Joint Koopman-based Learning of System Models and Function Dictionaries
A novel method is presented that jointly learns a function dictionary and lifted bilinear model purely from data by incorporating the Koopman model in a neural network architecture and achieves similar performance as a nonlinear MPC based on a nominal dynamics model in medium altitude.
Nonlinear System Level Synthesis for Polynomial Dynamical Systems
—This work introduces a controller synthesis method via system level synthesis for nonlinear systems char-acterized by polynomial dynamics. The resulting framework yields finite impulse response,
Model predictive control for robust quantum state preparation
A critical engineering challenge in quantum technology is the accurate control of quantum dynamics. Model-based methods for optimal control have been shown to be highly effective when theory and


Modeling and Control of Soft Robots Using the Koopman Operator and Model Predictive Control
This work describes this Koopman-based system identification method and its application to model predictive controller design, which yields an explicit control-oriented linear model rather than just a "black-box" input-output mapping.
Episodic Koopman Learning of Nonlinear Robot Dynamics with Application to Fast Multirotor Landing
A novel episodic method to learn a robot’s nonlinear dynamics model and an increasingly optimal control sequence for a set of tasks based on the Koopman operator approach to nonlinear dynamical systems analysis.
Data-driven feedback stabilization of nonlinear systems: Koopman-based model predictive control
The proposed feedback control design remains completely data-driven and does not require any explicit knowledge of the original system, and due to the bilinear structure of the Koopman model, seeking a CLF is no longer a bottleneck for LMPC.
Neural Lander: Stable Drone Landing Control Using Learned Dynamics
A novel deep-learning-based robust nonlinear controller (Neural-Lander) that improves control performance of a quadrotor during landing and is the first DNN-based nonlinear feedback controller with stability guarantees that can utilize arbitrarily large neural nets.
Episodic Learning with Control Lyapunov Functions for Uncertain Robotic Systems*
A machine learning framework centered around Control Lyapunov Functions to adapt to parametric uncertainty and unmodeled dynamics in general robotic systems and yields a stabilizing quadratic program model-based controller.
Feedback Stabilization Using Koopman Operator
The proposed controller design approach outlined in this paper is that it does not require explicit knowledge of system dynamics and the bilinear representation of a nonlinear control system in the Koopman eigenfunction space can be obtained from time-series data.
Linear predictors for nonlinear dynamical systems: Koopman operator meets model predictive control
Sparse identification of nonlinear dynamics for model predictive control in the low-data limit
The recent sparse identification of nonlinear dynamics (SINDY) modelling procedure is extended to include the effects of actuation and it is demonstrated that the resulting SINDY-MPC framework has higher performance, requires significantly less data, and is more computationally efficient and robust to noise than NN models, making it viable for online training and execution in response to rapid system changes.
Generalizing Koopman Theory to Allow for Inputs and Control
A new generalization of Koopman operator theory that incorporates the effects of inputs and control is developed that is rigorously connected and generalizes a recent development called Dynamic Mode Decomposition with control (DMDc).
Recent Advances in Quadratic Programming Algorithms for Nonlinear Model Predictive Control
The aim of this paper is to summarize the main algorithmic advances in the field and to provide a consistent benchmark between a selection of software tools that have been recently developed.