Prediction of Ship Motion Attitude Based on Combined Model

Abstract

Due to the influence of sea conditions, six dimensional movements, including heave, roll, pitch, sway, surge and yaw, are easy to be produced while ships sailing. These motions seriously affect the safety of its sailing, so the prediction of ship motion attitude is particularly important. In this case, a new combined model called CWGRU is proposed for predicting ship motion attitude with high accuracy. The CWGRU is based on complete ensemble empirical mode decomposition algorithm (CEEMD), whale optimization algorithm (WOA) and gated recurrent unit (GRU). Firstly, the CEEMD algorithm is used to decompose the ship’s sailing attitude data into a number of intrinsic mode functions (IMF) with different characteristics, so that the non-stationary time sequences have stability and periodicity. Then, the GRU based on WOA (WGRU) model is used to learn the short-term characteristics of each IMF component and predict it. Finally, the predicted values of each IMF component are added to obtain the prediction results. In order to verify the effectiveness of the CWGRU model proposed in this paper, the experiment based on real motion data collected in a ship are carried out. The first 80 of the data is used as the training set, and the last 20 is used for the test. Experimental results show that the performance of CWGRU is much better than that of GRU and WGRU.