基于深度学习的短时交通流预测方法综述与仿真研究  

Review and Simulation of Short-Term Traffic Flow Prediction Methods Based on Deep Learning

作　　者：朱仕威 叶宝林 吴维敏[3] ZHU Shiwei;YE Baolin;WU Weimin(School of Informatics Science and Technology,Zhejiang Sci-Tech University,Hangzhou 310018,China;College of Information Science and Engineering,Jiaxing University,Jiaxing 314001,China;State Key Laboratory of Industrial Control Technology,Zhejiang University,Hangzhou 310027,China)

机构地区：[1]浙江理工大学信息科学与工程学院,浙江杭州310018 [2]嘉兴学院信息科学与工程学院,浙江嘉兴314001 [3]浙江大学控制科学与工程学院工业控制技术国家重点实验室,浙江杭州310027

出　　处：《软件导刊》2024年第2期182-193,共12页Software Guide

基　　金：国家自然科学基金青年基金项目(61603154);浙江省自然科学基金探索项目(LTGS23F030002);浙江省自然科学基金一般项目(LY19F030014);工业控制技术国家重点实验室开放课题(ICT2022B52)。

摘　　要：近年来,随着城市路网交通检测设备和城市数据存储基础设施的升级换代以及深度学习技术的快速发展,应用深度学习技术解决城市路网短时交通流预测问题已成为智能交通领域的一个研究热点。不同于传统短时交通流预测方法,基于深度学习的短时交通流预测方法能充分利用海量交通流数据,深入挖掘路网中不同交通节点间流量的隐藏特征与复杂时空关联,能有效提升预测短时交通流的精度。首先,简要回顾短时交通流预测方法的发展历史,重点分析、讨论基于深度学习模型的短时交通流预测方法最新技术进展和理论研究结果。其次,梳理、总结国内外广泛用于验证算法有效性和进行比较分析的公开交通流数据集。再次,阐述基于深度学习模型的短时交通流预测算法解决实际交通流预测问题的具体过程和详细步骤,基于公开测试数据集PEMS04分别对基于深度学习模型长短时记忆网络(LSTM)和门控循环单元(GRU)的短时交通流预测算法进行仿真研究,以验证算法的有效性及其相较于传统方法的优势。最后,总结、展望基于深度学习模型的短时交通流预测方法在实际应用中存在的挑战和未来研究方向。In recent years,thanks to the upgrading of traffic detection equipment and urban data storage infrastructure as well as the rapid de⁃velopment of deep learning technology,the application of deep learning technology to solve the problem of short-term traffic flow prediction has become a research hotspot in the field of intelligent transportation.Different from the traditional short-time traffic flow prediction methods,the short-time traffic flow prediction method based on deep learning can make full use of massive traffic data,dig deeply the hidden features and associations between traffic nodes,and effectively improve the accuracy of short-term traffic flow prediction.Firstly,this paper briefly re⁃views the development history of short-term traffic flow prediction methods,and focuses on analyzing and discussing the latest technical prog⁃ress and theoretical research results of short-term traffic flow prediction methods based on deep learning model.Then,the open traffic flow da⁃ta sets,which are widely used to verify the effectiveness of the algorithm and make comparative analysis,are combed and summarized.In addi⁃tion,the specific process and detailed steps of applying the short-time traffic flow prediction algorithm based on the deep learning model to solve the actual traffic flow prediction problem are described. The short-term traffic flow prediction algorithm based on the deep learning modelLSTM and GRU is simulated with the open test data set PEMS04. Simulation results verify the effectiveness of the algorithm and its advantagescompared with traditional methods. Finally, the challenges and future development directions in the field of short-term traffic flow forecastinghave been summarized and prospected.

关 键 词：短时交通流预测 深度学习 时间序列 交通数据集 卷积神经网络 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]