Simulating Ship Manoeuvrability with Artificial Neural Networks Trained by a Short Noisy Data Set

Abstract

Artificial neural networks are applied to model the manoeuvrability characteristics of a ship based on empirical information acquired from experiments with a scaled model. This work aims to evaluate the performance of the proposed method of training the artificial neural network model even with a very small quantity of noisy data. The data used for the training consisted of zig-zag and circle manoeuvres carried out in agreement with the IMO standards. The wind effect is evident in some of the recorded experiments, creating additional disturbance to the fitting scheme. The method used for the training of the network is the Levenberg–Marquardt algorithm, and the results are compared with the scaled conjugate gradient method and the Bayesian regularization. The results obtained with the different methodologies show very suitable accuracy in the prediction of the referred manoeuvres.

Ship Maneuvering in Shallow and Narrow Waters: Predictive Methods and Model Development Review

“…Empirical regressions are based upon the most prevalent ship hull forms like single screw ships. Deviation from common hull forms, like in twin screw ships, can exceed the parametric range of the experimental database and potentially lead to inaccurate predictions [93][94][95][96].…”

“…Research into the automatic mooring of ships is considered one of the most complex problems in ship control. The ANN model has proven to be an extremely effective solution for the automatic docking of ships, as it can learn and imitate the actions of the human brain during docking maneuvers [96][97][98][99]. Neural network algorithms may have certain advantages since no structure of the mathematical model of the ship is required.…”

Prediction of Ship Main Particulars for Harbor Tugboats Using a Bayesian Network Model and Non-Linear Regression

“…Notably, its advantages include suitability for small and incomplete datasets, potential for structured learning, integration of diverse information sources, explicit treatment of uncertainty, and support for decision analysis and prompt responses. Leveraging these advantages, various Bayesian models have been demonstrated [28][29][30] for parameter identification and prediction of ship motions and maneuverability, contributing to the prediction of vessel hydrodynamics. Additionally, Bayesian networks find application in risk assessment [31][32][33][34], accident scenario analysis [35][36][37], reliability analysis [38,39], and fuel consumption analysis [40,41] within the maritime domain.…”

Maritime Autonomous Vessels

“…The main goal of this book is to address key challenges, thereby promoting research on marine autonomous ships. There are many topics on autonomous vessels involved in this book, for instance, automatic control [1][2][3][4], manoeuvrability [5][6][7][8], collision avoidance [9][10][11], ship target identification [12][13][14][15], motion planning [16], and buckling analysis [17].…”

“…The validation was carried out by comparing the result of the measured values with the predictions obtained using the manoeuvring models. Moreira and Guedes Soares [6] implemented ANNs to predict the heading angle and trajectories of a model ship from the output rudder angle command. The main feature of this study is that it demonstrates that the ANN can learn even from a short and noisy data set.…”