no code implementations • 2 Apr 2024 • Jan H. Hoekstra, Chris Verhoek, Roland Tóth, Maarten Schoukens
This model structure is able to represent many common model augmentation structures, thus unifying them under the proposed model structure.
no code implementations • 5 Jan 2024 • Jonas Weigand, Gerben I. Beintema, Jonas Ulmen, Daniel Görges, Roland Tóth, Maarten Schoukens, Martin Ruskowski
However, in continuous-time state-space model estimation, it has been observed that improper normalization of either the hidden state or hidden state derivative of the model estimate, or even of the time interval can lead to numerical and optimization challenges with deep learning based methods.
1 code implementation • 13 Jul 2023 • Gerben I. Beintema, Maarten Schoukens, Roland Tóth
Available methods for identification of stochastic dynamical systems from input-output data generally impose restricting structural assumptions on either the noise structure in the data-generating system or the possible state probability distributions.
no code implementations • 2 May 2023 • Sarvin Moradi, Nick Jaensson, Roland Tóth, Maarten Schoukens
Hamiltonian Neural Networks (HNNs) implement Hamiltonian theory in deep learning and form a comprehensive framework for modeling autonomous energy-conservative systems.
no code implementations • 4 Apr 2023 • Rishi Ramkannan, Gerben I. Beintema, Roland Tóth, Maarten Schoukens
The SUBNET neural network architecture has been developed to identify nonlinear state-space models from input-output data.
no code implementations • 30 Mar 2023 • Jan H. Hoekstra, Bence Cseppentő, Gerben I. Beintema, Maarten Schoukens, Zsolt Kollár, Roland Tóth
Artificial neural networks (ANN) have been shown to be flexible and effective function estimators for identification of nonlinear state-space models.
no code implementations • 16 Jan 2023 • Lucian Cristian Iacob, Maarten Schoukens, Roland Tóth
The Koopman framework is a popular approach to transform a finite dimensional nonlinear system into an infinite dimensional, but linear model through a lifting process, using so-called observable functions.
no code implementations • 27 Nov 2022 • David van de Sanden, Maarten Schoukens, Mauro Salazar
Our numerical results show that the proposed data-driven pricing scheme can effectively align the users' flows with the system optimum, significantly reducing the societal costs with respect to the uncontrolled flows (by about 15% and 25% depending on the scenario), and respond to environmental changes in a robust and efficient manner.
no code implementations • 26 Oct 2022 • Gerben I. Beintema, Maarten Schoukens, Roland Tóth
In this paper, we aim to overcome these issues by proposing a method which uses a truncated prediction loss and a subspace encoder for state estimation.
no code implementations • 25 Jul 2022 • Lucian Cristian Iacob, Roland Tóth, Maarten Schoukens
In applications for systems with inputs, generally a linear time invariant (LTI) form of the Koopman model is assumed, as it facilitates the use of control techniques such as linear quadratic regulation and model predictive control.
no code implementations • 15 Jun 2022 • Lucian Cristian Iacob, Roland Tóth, Maarten Schoukens
In the lifted space, the dynamics are linear and represented by a so-called Koopman operator.
1 code implementation • 20 Apr 2022 • Gerben I. Beintema, Maarten Schoukens, Roland Tóth
Continuous-time (CT) modeling has proven to provide improved sample efficiency and interpretability in learning the dynamical behavior of physical systems compared to discrete-time (DT) models.
no code implementations • 8 Apr 2022 • Chris Verhoek, Gerben I. Beintema, Sofie Haesaert, Maarten Schoukens, Roland Tó th
The Linear Parameter-Varying (LPV) framework provides a modeling and control design toolchain to address nonlinear (NL) system behavior via linear surrogate models.
1 code implementation • 2 Apr 2022 • Wouter Kouw, Albert Podusenko, Magnus Koudahl, Maarten Schoukens
We propose a variational Bayesian inference procedure for online nonlinear system identification.
no code implementations • 6 Oct 2021 • Lucian Cristian Iacob, Gerben Izaak Beintema, Maarten Schoukens, Roland Tóth
The present paper treats the identification of nonlinear dynamical systems using Koopman-based deep state-space encoders.
no code implementations • 26 Mar 2021 • Maarten Schoukens
The identification of black-box nonlinear state-space models requires a flexible representation of the state and output equation.
1 code implementation • 14 Dec 2020 • Gerben Beintema, Roland Toth, Maarten Schoukens
This paper introduces a method that approximates the simulation loss by splitting the data set into multiple independent sections similar to the multiple shooting method.
no code implementations • 14 Dec 2020 • Thiago B. Burghi, Maarten Schoukens, Rodolphe Sepulchre
After sixty years of quantitative biophysical modeling of neurons, the identification of neuronal dynamics from input-output data remains a challenging problem, primarily due to the inherently nonlinear nature of excitable behaviors.
2 code implementations • 14 Dec 2020 • Gerben Izaak Beintema, Roland Toth, Maarten Schoukens
An encoder function, represented by a neural network, is introduced to learn a reconstructability map to estimate the model states from past inputs and outputs.
no code implementations • 22 Feb 2020 • Thiago B. Burghi, Maarten Schoukens, Rodolphe Sepulchre
This paper applies the classical prediction error method (PEM) to the estimation of nonlinear discrete-time models of neuronal systems subject to input-additive noise.
no code implementations • 15 Jan 2020 • Dhruv Khandelwal, Maarten Schoukens, Roland Tóth
Model structure and complexity selection remains a challenging problem in system identification, especially for parametric non-linear models.
no code implementations • 5 Apr 2019 • Dhruv Khandelwal, Maarten Schoukens, Roland Tóth
Based on the results achieved for the case studies, we critically analyse the performance of the proposed method.