|
|
Publications
An, G.S., Xiang, G., Xiang, X.B. and Guedes Soares, C. (2024), System identification method of ship manoeuvring motion driven by knowledge and data, Advances in Maritime Technology and Engineering, Guedes Soares, C. & Santos T.A. (Eds.), Taylor and Francis Group, London, UK¸ pp. 295-303.
This paper proposes a machine learning ship motion system identification method using a Physics-Informed Neural Network (PINN) driven by knowledge and data. The ship motion model with coefficients is incorporated into the loss function as a regularization term, enabling the PINN to conform to training data distribution and adhere to differential equations governing physical laws. Research shows that PINN can learn models with improved generalization and prediction accuracy using fewer data samples, furthermore, they exhibit high robustness. The techniques and discoveries of this study can furnish theoretical direction and methodological backing for the precise modeling of ship manoeuvring motion, thereby pioneering a novel approach in system identification theory that merges knowledge and data-driven approaches.
If you did not manage to obtain a copy of this paper: Request a copy of this article
For information about all CENTEC publications you can download: Download the Complete List of CENTEC Publications
|