Neural network based combining prediction model and its application in ship motion prediction

The single-step neural network combining prediction model and the multi-step neural network combining prediction model are researched on this paper, and those models are applied to the ship pitch motion prediction. Firstly the single-step neural network combination prediction model is proposed and the optimal combination of the conventional single-step prediction model attained by the neural network. And based on the single-step neural network combination prediction model, the multistep neural network prediction model is proposed, and we get the optimal combination of the conventional single-step prediction model by the neural network. Used the periodgram model and AR model combined by the two neural networks model, the ship pitch motion data are predicted. The simulation result shows that these two combining models are more accurate than the single prediction model and the multi-step combining model can get better results in phase tracking and the error of amplitude.