The Ship Motion Attitude Prediction Based on VMD-TCN Model

Solving the problem of ship motion attitude prediction at sea will bring many benefits, the traditional prediction methods and single prediction model often have large errors. In this paper, a combined prediction model based on variational mode decomposition(VMD) and temporal convolutional network(TCN) is proposed. First, use VMD to decompose the ship roll and pitch angle data into several intrinsic mode functions (IMFs), then using the TCN to predict attitude component, and finally the prediction results are added together to get the prediction value of the ship attitude. The simulation result demonstrate that TCN can accurately predict the ship attitude, and its improved model VMD-TCN has substantially improved the accuracy than the TCN model, It proves that the model in this paper is effective in ship attitude prediction.