摘要
在内河通航船舶数量不断增加和通航环境日益复杂的情况下,为充分挖掘海量AIS数据中的价值信息,针对内河船舶航迹预测中的精度和可靠性问题,应用循环神经网络方法,提出了一种结合深度学习注意力机制(Attention Mechanism)和长短期记忆网络(Long Short-Term Memory,LSTM)的船舶航迹预测模型。该模型以历史时刻的船舶位置、速度及航向等数据为基础,考虑AIS数据与船舶航迹的时间序列特性,基于LSTM编码-解码航迹预测基本模型,通过引入时间与空间注意力机制,模拟了船舶自身航行模式和船舶交互作用对航迹预测的影响,定义了模型的损失函数和输出方式,构建了完整的Atten-LSTM航迹预测模型。应用海事AIS数据进行模型训练和航迹预测分析,实验结果表明在船舶安全航行条件下,Atten-LSTM模型具有易实现、精度高、可靠性强的特点。
A method of ship track prediction based on combination of attention mechanism and long short-term memory is developed to cope with the complexity of inland river traffic that brought about by navigational environment and heavy traffic density.Valuable information is extracted from AIS data mining with the method.A LSTM encode-decode prediction model is built for process ship historical AIS data according to the sequential character of AIS position data.The temporal and spatial attention mechanism is introduced into the basic LSTM encode-decode prediction model to simulate the motion of ships sailing on their own and the impact of encountering situations.The loss function and output mode are defined and the attention-LSTM model of recurrent neural network type is completed.The model is trained and verified with existing AIS data sets.
作者
刘成勇
乔文杰
陈蜀喆
万一
LIU Chengyong;QIAO Wenjie;CHEN Shuzhe;WAN Yi(School of Navigation,Wuhan University of Technology,Wuhan 430063,China;Hubei Inland Technology Key Laboratory,Wuhan 430063,China)
出处
《中国航海》
CSCD
北大核心
2021年第4期94-100,106,共8页
Navigation of China
基金
国家自然科学基金青年科学基金项目(51809207)
作者简介
刘成勇(1976-),男,湖北武汉人,副教授,工学博士,研究方向为交通环境与安全保障。E-mail:lcywhut@163.com
