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

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