RECURSIVE NEURAL NETWORK MODEL OF CATAMARAN MANOEUVRING

Abstract

A neural network model to simulate catamaran manoeuvres is proposed as an alternative to the traditional methodology of developing manoeuvring mathematical models. Data obtained in full-scale trials with a real ship are used to train the model. By recording full-scale trials of catamaran manoeuvres it is possible to generate a neural network model which will allow the prediction of the catamaran manoeuvring performance under different conditions.

A Recursive Neural Network (RNN) manoeuvring simulation model is proposed and applied to a catamaran in this specific case. Inputs to the simulation are the orders of rudder angle and ship’s speed and also the recursive outputs velocities of sway and yaw. Two types of manoeuvres are simulated: tactical circles and zigzags. The results between the full-scale data and the simulations are compared in order to analyze and determine the accuracy of the RNN. The study is performed for a catamaran operating in the Tagus estuary for passenger transport to and from Lisbon.