Details
| Original language | English |
|---|---|
| Title of host publication | Proceedings of the 21st International Conference on Informatics in Control, Automation and Robotics |
| Pages | 55-66 |
| Number of pages | 12 |
| Volume | 1 |
| ISBN (electronic) | 978-989-758-717-7 |
| Publication status | Published - 2024 |
Publication series
| Name | Proceedings of the International Conference on Informatics in Control, Automation and Robotics |
|---|---|
| ISSN (Print) | 2184-2809 |
Abstract
Physics-informed neural networks (PINNs) are trained using physical equations and can also incorporate un modeled effects by learning from data. PINNs for control (PINCs) of dynamical systems are gaining interest due to their prediction speed compared to classical numerical integration methods for nonlinear state-space models, making them suitable for real-time control applications. We introduce the domain-decoupled physics informed neural network (DD-PINN) to address current limitations of PINC in handling large and complex nonlinear dynamical systems. The time domain is decoupled from the feed-forward neural network to con struct an Ansatz function, allowing for calculation of gradients in closed form. This approach significantly re duces training times, especially for large dynamical systems, compared to PINC, which relies on graph-based automatic differentiation. Additionally, the DD-PINN inherently fulfills the initial condition and supports higher-order excitation inputs, simplifying the training process and enabling improved prediction accuracy. Validation on three systems– a nonlinear mass-spring-damper, a five-mass-chain, and a two-link robot demonstrates that the DD-PINN achieves significantly shorter training times. In cases where the PINC’s pre diction diverges, the DD-PINN’s prediction remains stable and accurate due to higher physics loss reduction or use of a higher-order excitation input. The DD-PINN allows for fast and accurate learning of large dynamical systems previously out of reach for the PINC.
Keywords
- Model Learning, Physics-Informed Machine Learning, Surrogate Modeling, System Dynamics
ASJC Scopus subject areas
- Computer Science(all)
- Artificial Intelligence
- Computer Science(all)
- Signal Processing
Cite this
- Standard
- Harvard
- Apa
- Vancouver
- BibTeX
- RIS
Proceedings of the 21st International Conference on Informatics in Control, Automation and Robotics. Vol. 1 2024. p. 55-66 (Proceedings of the International Conference on Informatics in Control, Automation and Robotics).
Research output: Chapter in book/report/conference proceeding › Conference contribution › Research › peer review
}
TY - GEN
T1 - Domain-Decoupled Physics-informed Neural Networks with Closed-Form Gradients for Fast Model Learning of Dynamical Systems
AU - Krauss, Henrik
AU - Habich, Tim-Lukas
AU - Bartholdt, Max
AU - Seel, Thomas
AU - Schappler, Moritz
N1 - Publisher Copyright: © 2024 by SCITEPRESS– Science and Technology Publications, Lda.
PY - 2024
Y1 - 2024
N2 - Physics-informed neural networks (PINNs) are trained using physical equations and can also incorporate un modeled effects by learning from data. PINNs for control (PINCs) of dynamical systems are gaining interest due to their prediction speed compared to classical numerical integration methods for nonlinear state-space models, making them suitable for real-time control applications. We introduce the domain-decoupled physics informed neural network (DD-PINN) to address current limitations of PINC in handling large and complex nonlinear dynamical systems. The time domain is decoupled from the feed-forward neural network to con struct an Ansatz function, allowing for calculation of gradients in closed form. This approach significantly re duces training times, especially for large dynamical systems, compared to PINC, which relies on graph-based automatic differentiation. Additionally, the DD-PINN inherently fulfills the initial condition and supports higher-order excitation inputs, simplifying the training process and enabling improved prediction accuracy. Validation on three systems– a nonlinear mass-spring-damper, a five-mass-chain, and a two-link robot demonstrates that the DD-PINN achieves significantly shorter training times. In cases where the PINC’s pre diction diverges, the DD-PINN’s prediction remains stable and accurate due to higher physics loss reduction or use of a higher-order excitation input. The DD-PINN allows for fast and accurate learning of large dynamical systems previously out of reach for the PINC.
AB - Physics-informed neural networks (PINNs) are trained using physical equations and can also incorporate un modeled effects by learning from data. PINNs for control (PINCs) of dynamical systems are gaining interest due to their prediction speed compared to classical numerical integration methods for nonlinear state-space models, making them suitable for real-time control applications. We introduce the domain-decoupled physics informed neural network (DD-PINN) to address current limitations of PINC in handling large and complex nonlinear dynamical systems. The time domain is decoupled from the feed-forward neural network to con struct an Ansatz function, allowing for calculation of gradients in closed form. This approach significantly re duces training times, especially for large dynamical systems, compared to PINC, which relies on graph-based automatic differentiation. Additionally, the DD-PINN inherently fulfills the initial condition and supports higher-order excitation inputs, simplifying the training process and enabling improved prediction accuracy. Validation on three systems– a nonlinear mass-spring-damper, a five-mass-chain, and a two-link robot demonstrates that the DD-PINN achieves significantly shorter training times. In cases where the PINC’s pre diction diverges, the DD-PINN’s prediction remains stable and accurate due to higher physics loss reduction or use of a higher-order excitation input. The DD-PINN allows for fast and accurate learning of large dynamical systems previously out of reach for the PINC.
KW - Model Learning
KW - Physics-Informed Machine Learning
KW - Surrogate Modeling
KW - System Dynamics
UR - http://www.scopus.com/inward/record.url?scp=105001299316&partnerID=8YFLogxK
U2 - 10.5220/0012935200003822
DO - 10.5220/0012935200003822
M3 - Conference contribution
VL - 1
T3 - Proceedings of the International Conference on Informatics in Control, Automation and Robotics
SP - 55
EP - 66
BT - Proceedings of the 21st International Conference on Informatics in Control, Automation and Robotics
ER -