On neural network identification for low-speed ship maneuvering model

• Published in: Journal of marine science and technology 2022-03, Vol.27 (1), p.772-785
• Main Authors: Wakita, Kouki, Maki, Atsuo, Umeda, Naoya, Miyauchi, Yoshiki, Shimoji, Tohga, Rachman, Dimas M., Akimoto, Youhei
• Format: Article
• Language: English
• Subjects: Automatic control; Automotive Engineering; Berthing; Computational fluid dynamics; Computer applications; Control systems; Dynamic models; Engineering; Engineering Design; Engineering Fluid Dynamics; Fluid dynamics; Hydrodynamics; Identification; Identification methods; Low speed; Mechanical Engineering; Model testing; Modelling; Neural networks; Offshore Engineering; Original Article; Recurrent neural networks; Scale models; Ship handling; System identification; Tests
• ISSN: 0948-4280; 1437-8213
• DOI: 10.1007/s00773-021-00867-1

Several studies on ship maneuvering models have been conducted using captive model tests or computational fluid dynamics (CFD) and physical models, such as the maneuvering modeling group (MMG) model. A new system identification method for generating a low-speed maneuvering model using recurrent neural networks (RNNs) and free running model tests is proposed in this study. We especially focus on a low-speed maneuver such as the final phase in berthing to achieve automatic berthing control. Accurate dynamic modeling with minimum modeling error is highly desired to establish a model-based control system. We propose a new loss function that reduces the effect of the noise included in the training data. Besides, we revealed the following facts—an RNN that ignores the memory before a certain time improved the prediction accuracy compared with the “standard” RNN, and the manual random maneuver test was effective in obtaining an accurate berthing maneuver model. In addition, several low-speed free running model tests were performed for the scale model of the M.V. Esso Osaka. As a result, this paper showed that the proposed method using a neural network model could accurately represent low-speed maneuvering motions.