Short term Vessel Motion Forecasting Based on Wavelet Neural Network for Wave Feed-forward Dynamic Positioning

摘要

Within the development project of a new wave feed-forward DynamicrnPositioning (DP) system for pipelaying vessels, several techniques forrnan accurate short term forecasting of sea surface elevation and vesselrnmotion have been investigated. In the present paper, an artificialrnintelligence hybrid method is studied, consisting of waveletrndecomposition (WT) coupled with artificial neural networks (ANN)rnforecasting algorithms. The prediction for the three first order vesselrnmotions is performed. Forecasting performances of the proposed hybridrnmethod are compared with those of a simple ANN, showing that thernhybrid method produces better prediction.