计算机科学

计算机科学 ›› 2023, Vol. 50 ›› Issue (11A): 230800067-7.doi: 10.11896/jsjkx.230800067

• 大数据&数据科学 • 上一篇    下一篇

基于注意力机制和ConvLSTM的船舶交通流量预测算法

李刚1, 宋文2, 陈致远1   

  1. 1 国能黄骅港务有限责任公司 河北 沧州 061113
    2 山东大学海洋研究院 山东 青岛 266237
  • 发布日期:2023-11-09
  • 通讯作者: 李刚(gang.li.gg@chnenergy.com.cn)
  • 基金资助:
    国家自然科学基金(62102228);山东省自然科学基金(ZR2021QF063);交通运输部交通强国建设试点任务项目

Ship Traffic Flow Prediction Algorithm Based on Attention Mechanism and ConvLSTM

LI Gang1, SONG Wen2, CHEN Zhiyuan1   

  1. 1 CHN ENERGY Huanghua Port Affairs Co.,LTD,Cangzhou,Hebei 061113,China
    2 Insitute of Marine Science and Technology,Shandong University,Qingdao,Shandong 266237,China
  • Published:2023-11-09
  • About author:LI Gang,born in 1976,postgraduate,senior engineer.His main research interests include smart port construction,information and intelligence.
  • Supported by:
    National Natural Science Foundation of China(62102228),Shandong Provincial Natural Science Foundation(ZR2021QF063) and Pilot Task Project of the Ministry of Transport for the Construction of a Strong Transportation Country.

PDF (PC)

807

摘要： 船舶交通流量预测是港口智能交通系统的关键技术之一,对港口运输的效率和安全起到至关重要的作用。针对现有预测方法难以有效提取船舶交通流量数据中的时空相关性特征信息的问题,提出了一种基于注意力机制和ConvLSTM的预测方法(ACLN)。ACLN首先通过深度的ConvLSTM构建编码网络,有效提取船舶交通流量数据中的时空相关性特征信息;其次通过注意力机制对提取的时空相关特征信息进行重要性关注,使模型在预测的过程中自动关注更重要的特征信息;最后通过多层的ConvLSTM和CNN构建预测网络,对提取的时空特征信息进行解析并输出预测结果。在真实的港口船舶交通流量数据上验证了所提方法的有效性,实验结果表明,所提方法的预测性能明显优于目前公开的预测方法,能够对一定区域进行有效的长短时预测,具有一定的实用价值。

关键词: 智能交通, 智能港口, 船舶流量预测, 时空特征提取, 卷积神经网络, 长短期记忆网络, 卷积长短期记忆网络, 注意力机制

Abstract: Ship traffic flow prediction is one of the key technologies of port intelligent transportation system,which plays a vital role in the efficiency and safety of port transportation.Aiming at the problem that the existing prediction methods are difficult to effectively extract the spatio-temporal feature information from the ship traffic flow data,a prediction method based on attention mechanism and ConvLSTM(ACLN) is proposed.ACLN first constructs a encoding network through the deep ConvLSTM to effectively extract the spatio-temporal feature information from the ship traffic flow data.Secondly,the attention mechanism is used to pay attention to the importance of the extracted spatio-temporal feature information,so that the model can automatically pay attention to the more important feature information in the process of prediction.Finally,the prediction network is constructed by multiple layers of ConvLSTM and CNN to parse the extracted spatiotemporal feature information and output the prediction result.The effectiveness of the proposed method is verified on the real port ship traffic flow data.Experimental results show that the prediction performance of the proposed method is significantly better than that of the state-of-art prediction methods,and it can perform the long and short time prediction effectively in a certain area,and has a certain practical value.

Key words: Intelligent transportation, Intelligent port, Ship flow prediction, Spatio-temporal feature extraction, Convolutional neural network, Long short-term memory network, Convolutional long short-term memory network, Attention mechanism

中图分类号: 

  • TP393
[1]YAN X P,LIU C G.Review and prospect for intelligent waterway transportation system [J].CAAI Transactions on Intelligent Systems,2016,11(6):807-817.
[2]SUO Y F,CHEN W K,YANG S H,et al.Prediction of ship traf-fic flow based on deep neural network [J].Journal of Jimei University(Natural Science),2020,25(6):430-436.
[3]HE W,ZHONG C,SOTELO M A,et al.Short-term vessel traffic flow forecasting by using an improved Kalman model [J].Cluster Computing,2019,22(S4):7907-7916.
[4]WANG K,LIU W,CHEN J W,et al.Study on method of ship traffic flow prediction based on VMD and LSTM [J].Journal of Wuhan University of Technology(Transportation Science & Engineering),2022,46(1):177-182.
[5]LI X L,XIAO J L,LIU M J.Vessel traffic flow prediction based on the SARIMA model [J].Journal of Wuhan University of Technology(Transportation Science & Engineering),2017,41(2):329-332,337.
[6]GUO J,HUANG W,WILLIAMS B M.Adaptive Kalman filter approach for stochastic short-term traffic flow rate prediction and uncertainty quantification [J].Transportation Research Part C:Emerging Technologies,2014,43:50-64.
[7]WANG R F,CHEN Z Q,LIU Z X.Link prediction in complex networks with syncretic naive Bayes methods [J].CAAI Tran-sactions on Intelligent Systems,2019,14(1):99-107.
[8]YANG S S,WU C S,LIU Z,et al.Vessel traffic flow prediction using non-convex low-rank and sparsity constraints [J].Application Research of Computers,2018,35(1):43-47.
[9]YAN Z,YU C C,HAN L,et al.Short-term traffic flow forecasting method based on CNN+LSTM [J].Computer Engineering and Design,2019,40(9):2620-2624,2659.
[10]HUANG W,SONG G,HONG H,et al.Deep architecture fortraffic flow prediction:deep belief networks with multitask learning [J].IEEE Transactions on Intelligent Transportation Systems,2014,15(5):2191-2201.
[11]LAI J H,LIANG S.Short-time traffic flow forecasting method based on BP neural network optimized by ACS [J].Computer Engineering and Applications,2014,50(10):244-248.
[12]LI M,HAN D,WANG W.Vessel traffic flow forecasting by RSVR with chaotic cloud simulated annealing genetic algorithm and KPCA [J].Neurocomputing,2015,157:243-255.
[13]XIONG T,QI Y,ZHANG W B,et al.Short term traffic flow forecasting method based on temporal-spatial correlation [J].Computer Engineering and Design,2019,40(2):501-507.
[14]DENG S,JIA S,CHEN J.Exploring spatial-temporal relationsvia deep convolutional neural networks for traffic flow prediction with incomplete data [J].Applied Soft Computing,2019,78:712-721.
[15]ZHANG W,WU P,PENG Y,et al.Roll motion prediction ofunmanned surface vehicle based on coupled CNN and LSTM [J].Future Internet,2019,11(11):243.
[16]WANG X,LI J,ZHANG T.A machine-learning model for zonal ship flow prediction using AIS Data:a case study in the south atlantic states region [J].Journal of Marine Science and Engineering.2019,7(12):463.
[17]LIU J X,LIU Z D,ZHOU F.A marine traffic flow forecasting model based on generalized regression neural network [J].Navigation of China,2011,34(2):74-77.
[18]LI Q Y,JIN J,WANG B.Automatic sleep staging model based on the bi-directional LSTM convolutional network and attention mechanism [J].CAAI Transactions on Intelligent Systems,2022,17(3):523-530.
[19]HAO S,LEE D,ZHAO D.Sequence to sequence learning with attention mechanism for short-term passenger flow prediction in large-scale metro system [J].Transportation Research Part C:Emerging Technologies,2019,107:287-300.
[20]SHI X,CHEN Z,WANG H,et al.Convolutional LSTM network:a machine learning approach for Precipitation nowcasting [J].Advances in Neural Information Proces-sing Systems,2015,28:802-810.
[21]MNIH V,HEESS H,GRAVES A,et al.Recurrent models of visual attention [C]//Advances in Neural Information Proces-sing Systems,2014.
[22]KALTEH A M.Monthly river flow forecasting using artificial neural network and support vector regression models coupled with wavelet transform [J].Computers & Geosciences,2013,54:1-8.
[23]XIE Z,LIU Q.LSTM networks for vessel traffic flow prediction in inland waterway [C]//IEEE International Conference on Big Data and Smart Computing(BigComp).2018:418-425.
[24]ZHOU X,LIU Z,WANG F,et al.Using deep learning to forecast maritime vessel flows [J].Sensors,2020,20(6):1761.
Viewed
Full text


Abstract

Cited
  Shared   
  Discussed   
No Suggested Reading articles found!
渝ICP备16013121号-4
地址：重庆市渝北区洪湖西路18号    邮编：401121  
电话：023-63500828（编辑部） 023-67039612（会议/专题） 023-67039623（财务/发票）
E-mail：jsjkx12@163.com
本系统由北京玛格泰克科技发展有限公司设计开发