IEEE International Conference on Robotics and Automation - ICRA 2025, Atlanta (United States of America). 19-23 May 2025
Summary:
Pneumatic Artificial Muscles (PAMs) are complex nonlinear systems characterized by hysteresis, making them challenging to model with classical system identification methods. While deep learning has emerged as a powerful tool for modeling nonlinear systems from data, purely neural network-based models often lack interpretability and are prone to overfitting. To address these challenges, this study explores several hybrid approaches that combine analytical models with neural networks to model PAM behavior more effectively. The results demonstrate that hybrid models significantly outperform both purely analytical and black-box neural network models, particularly in terms of generalization and dynamic accuracy. Among the approaches, the Physics-Informed Neural Network (PINN) unsupervised model shows the most robust performance, capturing complex PAM dynamics while maintaining computational efficiency. These findings suggest that hybrid modeling is a promising and scalable solution for accurately representing the intricate behavior of PAMs.
Spanish layman's summary:
Para representar la dinámica compleja de los Músculos Artificiales Neumáticos, las Redes Neuronales Informadas por la Física (PINN) destacan por su rendimiento superior. Estos modelos superan a los enfoques analíticos y a las redes neuronales puras en capacidad de generalización y precisión dinámica.
English layman's summary:
Physics-Informed Neural Networks (PINNs) and other hybrid models significantly improve the modeling of complex, hysteretic Pneumatic Artificial Muscle dynamics. Combining analytical and neural network approaches yields superior generalization and accuracy over purely analytical or black-box methods.
Keywords: Artificial muscles , Training , Analytical models , Computational modeling , Closed box , Artificial neural networks , Data models , System identification , Nonlinear systems , Overfitting
DOI: 
https://doi.org/10.1109/ICRA55743.2025.11127748
Published in: ICRA 2025: Conference proceedings, pp: 3407-3413, ISBN: 979-8-3315-4140-8
Publication date: 02-Sep-2025.
Citation:
G. Wang, R. Chalard, J. Cifuentes, M.T. Pham, "Physics-Informed Hybrid Modeling of Pneumatic Artificial Muscles", presented at IEEE International Conference on Robotics and Automation - ICRA 2025, Atlanta, United States of America, 19-23 May 2025. In: ICRA 2025: Conference proceedings, pp. 3407-3413, doi: 10.1109/ICRA55743.2025.11127748
IIT-25-366C
