多船会遇场景下基于循环神经网络的船舶航速预测
Ship speed prediction based on recurrent neural network in multi-ship encounter scenarios
严忠伟 1赵建森 1吴欣雨 1王胜正 1陈信强 2高原1
作者信息
- 1. 上海海事大学商船学院,上海 201306
- 2. 上海海事大学物流科学与工程研究院,上海 201306
- 折叠
摘要
为进一步提高复杂环境下的船舶航速预测精度,提出一种在多船会遇场景下基于循环神经网络(recurrent neural network,RNN)的 船舶航速预测 模型.从船舶自动识 别系统(automatic identification system,AIS)数据中提取构成多船会遇场景的船舶航行动态信息(时间、航速等),采用插值法进行等时间间隔化处理,并构建基于RNN的船舶航速预测模型.采用长江口外水域的AIS数据,分别在不同会遇场景下进行实例验证.实验结果表明:在案例1和案例2场景下,RNN模型预测结果的平均绝对误差、均方误差、均方根误差、平均绝对百分比误差均比长短期记忆神经网络模型和支持向量机模型的小,说明RNN模型的预测精度比其他两种模型的高.
Abstract
In order to further improve the accuracy of ship speed prediction in complex environments,a ship speed prediction model based on the recurrent neural network(RNN)in multi-ship encounter scenarios is proposed.The dynamic information of ship navigation(time,speed,etc.)that composes the multi-ship encounter scenarios is extracted from the automatic identification system(AIS)data,the equal time-interval processing is performed by the interpolation method,and a ship speed prediction model based on RNN is constructed.The AIS data in the outside waters of the Yangtze River estuary is employed for case verification in different encounter scenarios.The experimental results show that,in the scenarios of case 1 and case 2,the mean absolute error,mean square error,root mean square error and mean absolute percentage error of the prediction results of the RNN model are smaller than those of the long-short term memory model and the support vector machine model,which indicates that the prediction accuracy of the RNN model is higher than that of the other two models.
关键词
交通安全/智能船舶/航速预测/循环神经网络(RNN)/船舶自动识别系统(AIS)Key words
traffic safety/intelligent ship/ship speed prediction/recurrent neural network(RNN)/automatic identification system(AIS)引用本文复制引用
基金项目
国家重点研发计划(2019YFB1600605)
国家自然科学基金(52102397)
国家自然科学基金(52071200)
国家自然科学基金(51978069)
国家自然科学基金(5207223)
大学生创新创业训练项目(G20220102)
出版年
2024
NETL
NSTL
万方数据