Generalizing Koopman Theory to Allow for Inputs and Control

@article{Proctor2018GeneralizingKT,
  title={Generalizing Koopman Theory to Allow for Inputs and Control},
  author={Joshua L. Proctor and Steven L. Brunton and J. Nathan Kutz},
  journal={SIAM J. Appl. Dyn. Syst.},
  year={2018},
  volume={17},
  pages={909-930}
}
We develop a new generalization of Koopman operator theory that incorporates the effects of inputs and control. Koopman spectral analysis is a theoretical tool for the analysis of nonlinear dynamical systems. Moreover, Koopman is intimately connected to Dynamic Mode Decomposition (DMD), a method that discovers spatial-temporal coherent modes from data, connects local-linear analysis to nonlinear operator theory, and importantly creates an equation-free architecture allowing investigation of… 

Figures from this paper

A data-driven Koopman model predictive control framework for nonlinear flows
The Koopman operator theory is an increasingly popular formalism of dynamical systems theory which enables analysis and prediction of the nonlinear dynamics from measurement data. Building on the
Koopman Representations of Dynamic Systems with Control
TLDR
This work provides an analysis framework to evaluate and rank proposed simplifications to the general Koopman representation for controlled dynamical systems and presents a hybrid formulation which independently and jointly observes the state and control inputs.
Data-driven discovery of Koopman eigenfunctions for control
TLDR
This work illustrates a fundamental closure issue of this approach and argues that it is beneficial to first validate eigenfunctions and then construct reduced-order models in these validated eigen Functions, termed Koopman Reduced Order Nonlinear Identification and Control (KRONIC).
Koopman Operator Theory for Nonlinear Dynamic Modeling using Dynamic Mode Decomposition
TLDR
A brief summary of the Koopman operator theorem for nonlinear dynamics modeling is provided and several data-driven implementations using dynamical mode decomposition (DMD) for autonomous and controlled canonical problems are analyzed.
Data-Driven Model Predictive Control using Interpolated Koopman Generators
TLDR
It is shown that when using the Koopman generator, this relaxation --- realized by linear interpolation between two operators --- does not introduce any error for control affine systems, which allows us to control high-dimensional nonlinear systems using bilinear, low-dimensional surrogate models.
Persistence of Excitation for Koopman Operator Represented Dynamical Systems
TLDR
This work investigates persistence of excitation conditions for discrete-time nonlinear systems that can be represented using Koopman operators, and illustrates these concepts on a repressilator circuit, a classical biological system from the field of synthetic biology, to motivate the proper selection of initial conditions.
Koopman Operator Theory for Nonautonomous and Stochastic Systems
TLDR
This chapter states basic properties of the Koopman operator and compares performance of the DMD and Arnoldi-type algorithms in the context of both definitions, and introduces the associated stochastic Koop man operator family.
Decomposition of Nonlinear Dynamical Systems Using Koopman Gramians
TLDR
The notion of an input-Koopman operator is introduced, and it is shown how it can be used to cast a nonlinear system into the classical state-space form, and conditions under which input and state observable functions are well separated are identified.
Learning Koopman Operators for Systems with Isolated Critical Points
TLDR
This paper considers the convergence of discrete-time Koopman approximation error to the continuous-time error and shows both how this convergence occurs in general and how it can fail for systems with multiple isolated critical points.
...
...

References

SHOWING 1-10 OF 62 REFERENCES
Koopman Invariant Subspaces and Finite Linear Representations of Nonlinear Dynamical Systems for Control
TLDR
This work presents a data-driven strategy to identify relevant observable functions for Koopman analysis by leveraging a new algorithm to determine relevant terms in a dynamical system by ℓ1-regularized regression of the data in a nonlinear function space and demonstrates the usefulness of nonlinear observable subspaces in the design of Koop man operator optimal control laws for fully nonlinear systems using techniques from linear optimal control.
A Data–Driven Approximation of the Koopman Operator: Extending Dynamic Mode Decomposition
TLDR
This approach is an extension of dynamic mode decomposition (DMD), which has been used to approximate the Koopman eigenvalues and modes, and if the data provided to the method are generated by a Markov process instead of a deterministic dynamical system, the algorithm approximates the eigenfunctions of the Kolmogorov backward equation.
Dynamic Mode Decomposition with Control
TLDR
This work develops a new method which extends dynamic mode decomposition (DMD) to incorporate the effect of control to extract low-order models from high-dimensional, complex systems and provides the additional innovation of being able to disambiguate between the underlying dynamics and the effects of actuation, resulting in accurate input-output models.
Applied Koopmanism.
TLDR
The Koopman framework is showing potential for crossing over from academic and theoretical use to industrial practice, and the paper highlights its strengths in applied and numerical contexts.
Spectral analysis of nonlinear flows
We present a technique for describing the global behaviour of complex nonlinear flows by decomposing the flow into modes determined from spectral analysis of the Koopman operator, an
Analysis of Fluid Flows via Spectral Properties of the Koopman Operator
This article reviews theory and applications of Koopman modes in fluid mechanics. Koopman mode decomposition is based on the surprising fact, discovered in Mezic (2005), that normal modes of linear
Koopman-mode decomposition of the cylinder wake
  • S. Bagheri
  • Engineering
    Journal of Fluid Mechanics
  • 2013
Abstract The Koopman operator provides a powerful way of analysing nonlinear flow dynamics using linear techniques. The operator defines how observables evolve in time along a nonlinear flow
Spectral Properties of Dynamical Systems, Model Reduction and Decompositions
In this paper we discuss two issues related to model reduction of deterministic or stochastic processes. The first is the relationship of the spectral properties of the dynamics on the attractor of
Reduced-order models for control of fluids using the eigensystem realization algorithm
TLDR
This article shows that using the Eigensystem Realization Algorithm (ERA) one can theoretically obtain exactly the same reduced-order models as by balanced POD, and demonstrates them on an example of the flow past an inclined flat plate at a low Reynolds number.
...
...