Asymptotically Stable Data-Driven Koopman Operator Approximation with Inputs using Total Extended DMD

Louis Lortie,James Richard Forbes

2024-08-30

Abstract:The Koopman operator framework can be used to identify a data-driven model of a nonlinear system. Unfortunately, when the data is corrupted by noise, the identified model can be biased. Additionally, depending on the choice of lifting functions, the identified model can be unstable, even when the underlying system is asymptotically stable. This paper presents an approach to reduce the bias in an approximate Koopman model, and simultaneously ensure asymptotic stability, when using noisy data. Additionally, the proposed data-driven modeling approach is applicable to systems with inputs, such as a known forcing function or a control input. Specifically, bias is reduced by using a total least-squares, modified to accommodate inputs in addition to lifted inputs. To enforce asymptotic stability of the approximate Koopman model, linear matrix inequality constraints are augmented to the identification problem. The performance of the proposed method is then compared to the well-known extended dynamic mode decomposition method and to the newly introduced forward-backward extended dynamic mode decomposition method using a simulated Duffing oscillator dataset and experimental soft robot arm dataset.

Systems and Control

What problem does this paper attempt to address?

### Problems the Paper Attempts to Solve This paper aims to address the following two main issues: 1. **Reducing the Bias of Koopman Models Due to Noisy Data**: - When data is affected by noise, the Koopman model identified through this data may be biased. This bias can affect the dynamic matrix and input matrix of the model, leading to a shift in eigenvalues, which in turn affects the model's predictive performance. - To reduce this bias, the paper proposes a method based on Total Least Squares (TLS), called Total Extended Dynamic Mode Decomposition (TEDMD). This method considers noise not only in the dynamic matrix but also in the input matrix. 2. **Ensuring the Asymptotic Stability of the Approximate Koopman Model**: - Even if the underlying system is asymptotically stable, the identified Koopman model may be unstable if the chosen lifting functions are inappropriate. - To ensure that the identified Koopman model is asymptotically stable, the paper introduces Linear Matrix Inequality (LMI) constraints. These constraints are added to the optimization problem to enforce the model's asymptotic stability. ### Summary The paper proposes a new method called Total Extended Dynamic Mode Decomposition with Input and Asymptotic Stability (TEDMD-AS). This method can reduce model bias when dealing with noisy data and ensure the model's asymptotic stability. By comparing with existing methods (such as EDMD, fbEDMD, etc.), the paper verifies the effectiveness and superiority of the proposed method. Specific applications include the simulated Duffing oscillator dataset and the experimental soft robotic arm dataset.

Asymptotically Stable Data-Driven Koopman Operator Approximation with Inputs using Total Extended DMD

Forward-Backward Extended DMD with an Asymptotic Stability Constraint

A Data-Driven Approximation of the Koopman Operator: Extending Dynamic Mode Decomposition

Closed-Loop Koopman Operator Approximation

System norm regularization methods for Koopman operator approximation

Nonlinear Data-Driven Approximation of the Koopman Operator

Koopman Operator Theory for Nonlinear Dynamic Modeling using Dynamic Mode Decomposition

Data-Driven Nonlinear Stabilization Using Koopman Operator

Robust Approximation of the Stochastic Koopman Operator

Dual-Loop Robust Control of Biased Koopman Operator Model by Noisy Data of Nonlinear Systems

Robust data-driven control for nonlinear systems using the Koopman operator

Data-Driven Models for Control Engineering Applications Using the Koopman Operator

Derivative-Based Koopman Operators for Real-Time Control of Robotic Systems

Data-driven optimal control of unknown nonlinear dynamical systems using the Koopman operator

Kernel-based Koopman approximants for control: Flexible sampling, error analysis, and stability

Learning Stable Models for Prediction and Control

Control-Coherent Koopman Modeling: A Physical Modeling Approach

Multiplicative Dynamic Mode Decomposition

Auxiliary Functions as Koopman Observables: Data-Driven Analysis of Dynamical Systems via Polynomial Optimization

Direct Data-Driven Stabilization of Nonlinear Affine Systems Via the Koopman Operator.

Data-Driven Approximations of Dynamical Systems Operators for Control