A Recurrent Neural Network (RNN) model is presented in this paper to predict the ship maneuvering motion. Inputs to the model are the orders of rudder angle and its variation as well as the propeller speed (ship speed) and also the recursive outputs velocities of surge, sway and yaw. The past values for the velocities are retained in the inputs to indicate the influence of historical state of motion on the maneuvering prediction. The KRISO Container Ship (KCS) is taken as the study object. The data obtained from a manoeuvring mathematical model and free-running model test are respectively used to train the neural network. Tactical circles and zigzags are simulated by the RNN, the prediction for maneuvers not involved in the training set shows that the RNN in this paper has good generalization performance. The concept of uncertainty is proposed to be considered in the further work through the analysis.
Recurrent neural networks for nonparametric modeling of ship maneuvering motion
2022
Aufsatz (Zeitschrift)
Elektronische Ressource
Unbekannt
Metadata by DOAJ is licensed under CC BY-SA 1.0
Recurrent neural networks for nonparametric modeling of ship maneuvering motion
Elsevier | 2022
|Ship maneuvering system, ship, and ship maneuvering method
Europäisches Patentamt | 2023
|SHIP MANEUVERING SYSTEM, SHIP, AND SHIP MANEUVERING METHOD
Europäisches Patentamt | 2021
|SHIP MANEUVERING SYSTEM, SHIP, AND SHIP MANEUVERING METHOD
Europäisches Patentamt | 2018
|SHIP MANEUVERING SYSTEM, SHIP, AND SHIP MANEUVERING METHOD
Europäisches Patentamt | 2018
|