摘要
针对船舶航行行为多维度的特点和船舶交通服务系统(Vessel Traffic Service,VIS)对船舶行为预测的精确度和实时性需求,提出结合船舶自动识别系统(Automatic Identification System,AIS)信息和BP(Back Propagation)神经网络的船舶航行行为预测方法。构造基于AIS信息的船舶航行行为特征表达方法,根据BP神经网络预测的基本原理,以连续3个时刻的船舶航行行为特征值为输入,以第4个时刻的船舶航行行为特征值为输出,对BP神经网络进行训练,用于对未来船舶航行行为进行预测。以成山角VTS水域内的船舶AIS信息为例进行试验,结果表明:利用该方法对船舶航行行为特征值进行预测的结果准确、实时,误差在可接受的范围内。
In view of the multi-dimensional characteristics of vessel behavior, a novel method of vessel behavior prediction based on AIS（ Automatic Identification System） data and BP（ Back Propagation） neural network is proposed to satisfy the requirement of VTS（ Vessel Traffic Service ） for accurate real-time vessel behavior prediction. The feature expressions of vessel behavior based on AIS data is established. The training process is as following： the longitude, latitude, heading and speed of a vessel is taken as the ship behavior feature input to the neural network ; data at three consecutive times are input to the network, and the fourth data following the input is output to train the BP neural network. The trained BP neural network is applied to the prediction of vessel behavior. The effectiveness and capability of the proposed method is verified with the AIS data from the waters of Chengshanjiao VTS. The results show that the method can predict the characteristics of vessel behavior timely with acceptable accuracy.
作者
甄荣
金永兴
胡勤友
施朝健
王胜正
ZHEN Rong JIN Yongxing HU Qinyou SHI Chaojian WANG Shengzheng(Merchant Marine College, Shanghai Maritime University, Shanghai 201306, China)
出处
《中国航海》
CSCD
北大核心
2017年第2期6-10,共5页
Navigation of China
基金
国家自然科学基金(51379121)
国家留学基金委联合培养博士生项目(201608310093)
上海市科委地方院校能力建设项目(15590501600)
上海海事大学研究生创新基金(2016ycx077)
上海海事大学优秀博士学位论文培养项目(2017bxlp003)
关键词
水路运输
船舶行为
预测
AIS信息
BP神经网络
waterway transportation
vessel behavior
prediction
AIS data
BP neural network
作者简介
甄荣（1990-），男，内蒙古乌兰察布人，博士生，从事海上交通信息处理方向研究。E—mail：zrandsea@163．com
