The invention discloses a ship motion prediction method based on a cyclic Transform model, and the method comprises the steps: carrying out the preprocessing of ship motion time series data, carrying out the segmentation of the preprocessed ship motion time series data, carrying out the segmentation of the segmented ship motion time series data, and greatly reducing the memory resources of a computer; based on the established ship motion prediction model of the circulating Transform, inputting the ship motion prediction model of the circulating Transform to the ship motion prediction model of the circulating Transform, and obtaining a block data output embedded vector sequence; and finally, obtaining the predicted value of the rolling angle, the predicted value of the pitching angle and the predicted value of the heaving height of the motion of the ship so as to predict the motion of the ship. According to the method, the long-range dependency relationship and the complex dynamic mode of the ship attitude are captured by adopting the ship motion time sequence data, and the prediction precision is improved. The established circulation Transform ship motion prediction model has good generalization ability, can adapt to real-time prediction requirements, can generate accurate ship attitude prediction results in a short time, can effectively capture nonlinearity and randomness of ship motion, and can better adapt to changes of different types of ships and marine environments.
本发明公开了一种基于循环Transformer模型的船舶运动预测方法,通过对船舶运动时间序列数据进行预处理,并将预处理后的船舶运动时间序列数据进行分段,在将分段后的船舶运动时间序列数据序列分块,大大降低了计算机内存资源;进而基于所建立的循环Transformer的船舶运动预测模型,输入至所述循环Transformer的船舶运动预测模型,获取块数据输出嵌入向量序列;最终获取船舶运动横摇角度的预测值、纵摇角度的预测值和垂荡高度的预测值,以对船舶的运动进行预测。本发明通过采用船舶运动时间序列数据,捕捉船舶姿态的长程依赖关系和复杂动态模式,提高预测精度。所建立的循环Transformer的船舶运动预测模型具有良好的泛化能力,能适应实时预测需求,能在短时间内生成准确的船舶姿态预测结果,同时能够有效捕捉船舶运动的非线性和随机性,更好的适应不同类型船舶和海洋环境的变化。
Ship motion prediction method based on cyclic Transform model
基于循环Transformer模型的船舶运动预测方法
2024-10-18
Patent
Electronic Resource
Chinese
Conv-Bi-LSTM model-based ship heaving motion prediction system and prediction method
European Patent Office | 2023
|Ship motion attitude prediction method based on multiple combinations
European Patent Office | 2023
|Application of Discrete Wavelet Transform to Ship Motion Analysis
British Library Online Contents | 2014
|