Article(id=1210314383859782475, tenantId=1146029695717560320, journalId=1189621681917173762, issueId=1210314382282716156, articleNumber=null, orderNo=null, doi=10.19620/j.cnki.1000-3703.20230582, pmid=null, cstr=null, oa=null, hot=null, price=null, onlineType=0, articleFormat=0, articleType=null, articleTypeStr=research-article, receivedDate=null, receivedDateStr=null, revisedDate=null, revisedDateStr=null, acceptedDate=null, acceptedDateStr=null, onlineDate=1766491035665, onlineDateStr=2025-12-23, pubDate=1711209600000, pubDateStr=2024-03-24, doiRegisterDate=null, doiRegisterDateStr=null, onlineIssueDate=1766491035665, onlineIssueDateStr=2025-12-23, onlineJustAcceptDate=null, onlineJustAcceptDateStr=null, onlineFirstDate=null, onlineFirstDateStr=null, sourceXml=null, magXml=null, createTime=1766491035665, creator=13701087609, updateTime=1766491035665, updator=13701087609, issue=Issue{id=1210314382282716156, tenantId=1146029695717560320, journalId=1189621681917173762, year='2024', volume='', issue='3', pageStart='1', pageEnd='62', issueExtLink='null', onlineDate='null', pubDate='null', beforeIssueId=null, nextIssueId=null, price=null, status=1, issueComplete=1, articleOrder=1, issueType=-1, specialIssue=null, createTime=1766491035290, creator=13701087609, updateTime=1766493673751, updator=13701087609, preIssue=null, nextIssue=null, ext={EN=IssueExt(id=1210325448853156641, tenantId=1146029695717560320, journalId=1189621681917173762, issueId=1210314382282716156, language=EN, specialIssueTitle=, coverIllustrator=null, specialIssueEditor=, specialIssueAbout=), CN=IssueExt(id=1210325448853156642, tenantId=1146029695717560320, journalId=1189621681917173762, issueId=1210314382282716156, language=CN, specialIssueTitle=, coverIllustrator=null, specialIssueEditor=, specialIssueAbout=)}, issueFiles=null}, startPage=24, endPage=30, ext={EN=ArticleExt(id=1210314384899969869, articleId=1210314383859782475, tenantId=1146029695717560320, journalId=1189621681917173762, language=EN, title=Research on Vehicle Trajectory Prediction Based on Dynamic Graph Attention, columnId=null, journalTitle=Automobile Technology, columnName=null, runingTitle=null, highlight=null, articleAbstract=

In current research on vehicle trajectory prediction, the existing Graph Attention Network (GAT), which is based on a static attention mechanism, fails to effectively capture interactions between vehicles in complex road conditions. To address this issue, this paper proposed an Encoder-Decoder Dynamic Graph Attention Network (ED-DGAT) to predict future trajectories of highway vehicles. In this model, the encoding module incorporates a dynamic graph attention mechanism to learn spatial interactions among vehicles. Simultaneously, a simplified dynamic graph attention network is adopted to model the interdependencies of vehicle movements during the decoding phase. This paper evaluated the proposed algorithm using the NGSIM dataset and conducted comparative analysis with other models such as LSTM, Social-LSTM (S-LSTM), and CS-LSTM. The results show that the Root Mean Squared Error (RMSE) of predicted trajectory has been reduced by 25%, and the inference speed is 2.61 times of the CS-LSTM model.

, correspAuthors=null, authorNote=null, correspAuthorsNote=null, copyrightStatement=null, copyrightOwner=null, extLink=null, articleAbsUrl=null, sourceXml=null, magXml=null, pdfUrl=null, pdf=null, pdfFileSize=null, pdfExtLink=null, richHtmlUrl=null, mobilePdfUrl=null, reviewReport=null, pdfFirstPage=null, abstractGraph=null, abstractGraphContent=null, abstractVideo=null, citation=null, cebUrl=null, magXmlContent=null, mapNumber=null, authorCompany=null, fund=null, authors=null, authorsList=Xiaowei Chen, Xuanpeng Li, Weigong Zhang), CN=ArticleExt(id=1210314385810133851, articleId=1210314383859782475, tenantId=1146029695717560320, journalId=1189621681917173762, language=CN, title=基于动态图注意力的车辆轨迹预测研究*, columnId=0, journalTitle=汽车技术, columnName=, runingTitle=null, highlight=null, articleAbstract=

针对目前轨迹预测研究中交互建模方法使用的图注意力网络(GAT)为静态注意力,无法有效捕捉复杂道路场景中车辆间交互的问题,提出了一种基于编码器-解码器架构的动态图注意力网络(ED-DGAT)预测高速公路环境中运动车辆的未来轨迹。编码模块使用动态图注意力机制学习场景中车辆间的空间交互,采用状态简化动态图注意力网络建模解码阶段车辆运动的相互依赖,最后使用NGSIM数据集评估所提出的模型,并与长短时记忆(LSTM)、联合社交池化与长短时记忆(S-LSTM)、联合卷积社交池化与长短时记忆(CS-LSTM)算法模型进行对比分析,结果表明,预测轨迹的均方根误差(RMSE)降低了25%,且模型的推理速度为CS-LSTM模型的2.61倍。

, correspAuthors=null, authorNote=null, correspAuthorsNote=null, copyrightStatement=null, copyrightOwner=null, extLink=null, articleAbsUrl=null, sourceXml=iM0stQxGn19K27mfSgHwDQ==, magXml=MK0t224pnReMpZzaYL/FPg==, pdfUrl=null, pdf=OnvenUcMQpBPWWFG8jPFdQ==, pdfFileSize=880785, pdfExtLink=null, richHtmlUrl=null, mobilePdfUrl=null, reviewReport=null, pdfFirstPage=null, abstractGraph=FUpa/3k+0ittfHYQ1ldQrw==, abstractGraphContent=null, abstractVideo=null, citation=null, cebUrl=null, magXmlContent=Q5qZWLWxsOv6ujpglvuJhQ==, mapNumber=null, authorCompany=null, fund=null, authors=null, authorsList=陈晓伟, 李煊鹏, 张为公)}, authors=[Author(id=1211006372150383407, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, orderNo=0, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=null, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1211006372225880882, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, authorId=1211006372150383407, language=EN, stringName=Xiaowei Chen, firstName=Xiaowei, middleName=null, lastName=Chen, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=null, address=Southeast University, Nanjing 210096, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1211006372301378355, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, authorId=1211006372150383407, language=CN, stringName=陈晓伟, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=null, address=东南大学,南京 210096, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1211006372066497323, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, xref=null, ext=[AuthorCompanyExt(id=1211006372074885932, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, companyId=1211006372066497323, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=Southeast University, Nanjing 210096), AuthorCompanyExt(id=1211006372083274541, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, companyId=1211006372066497323, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=东南大学,南京 210096)])]), Author(id=1211006372364292917, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, orderNo=1, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=null, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1211006372427207479, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, authorId=1211006372364292917, language=EN, stringName=Xuanpeng Li, firstName=Xuanpeng, middleName=null, lastName=Li, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=null, address=Southeast University, Nanjing 210096, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1211006372506899256, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, authorId=1211006372364292917, language=CN, stringName=李煊鹏, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=null, address=东南大学,南京 210096, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1211006372066497323, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, xref=null, ext=[AuthorCompanyExt(id=1211006372074885932, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, companyId=1211006372066497323, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=Southeast University, Nanjing 210096), AuthorCompanyExt(id=1211006372083274541, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, companyId=1211006372066497323, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=东南大学,南京 210096)])]), Author(id=1211006372561425210, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, orderNo=2, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=null, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1211006372624339772, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, authorId=1211006372561425210, language=EN, stringName=Weigong Zhang, firstName=Weigong, middleName=null, lastName=Zhang, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=null, address=Southeast University, Nanjing 210096, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1211006372695642941, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, authorId=1211006372561425210, language=CN, stringName=张为公, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=null, address=东南大学,南京 210096, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1211006372066497323, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, xref=null, ext=[AuthorCompanyExt(id=1211006372074885932, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, companyId=1211006372066497323, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=Southeast University, Nanjing 210096), AuthorCompanyExt(id=1211006372083274541, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, companyId=1211006372066497323, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=东南大学,南京 210096)])])], keywords=[Keyword(id=1211006372808889150, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, language=EN, orderNo=1, keyword=Trajectory prediction), Keyword(id=1211006372863415103, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, language=EN, orderNo=2, keyword=Attention mechanism), Keyword(id=1211006374050403136, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, language=EN, orderNo=3, keyword=Graph neural networks), Keyword(id=1211006374130094913, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, language=EN, orderNo=4, keyword=Multi-objective interaction), Keyword(id=1211006374188815170, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, language=CN, orderNo=1, keyword=轨迹预测), Keyword(id=1211006374247535427, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, language=CN, orderNo=2, keyword=注意力机制), Keyword(id=1211006374310449988, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, language=CN, orderNo=3, keyword=图神经网络), Keyword(id=1211006374394336069, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, language=CN, orderNo=4, keyword=多目标交互)], refs=[Reference(id=1211006375719736150, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, doi=null, pmid=null, pmcid=null, year=2009, volume=null, issue=null, pageStart=417, pageEnd=422, url=null, language=null, rfNumber=[1], rfOrder=0, authorNames=AMMOUN S, NASHASHIBI F, Cluj-Napoca, journalName=2009 IEEE 5th International Conference on Intelligent Computer Communication and Processing, refType=null, unstructuredReference=AMMOUN S, NASHASHIBI F. Real Time Trajectory Prediction for Collision Risk Estimation between Vehicles[C]// 2009 IEEE 5th International Conference on Intelligent Computer Communication and Processing. Cluj-Napoca, Romania: IEEE, 2009: 417-422., articleTitle=Real Time Trajectory Prediction for Collision Risk Estimation between Vehicles, refAbstract=null), Reference(id=1211006375778456407, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, doi=null, pmid=null, pmcid=null, year=2011, volume=3, issue=4, pageStart=4, pageEnd=19, url=null, language=null, rfNumber=[2], rfOrder=1, authorNames=LAUGIER C, PAROMTCHIK I E, PERROLLAZ M, journalName=IEEE Intelligent Transportation Systems Magazine, refType=null, unstructuredReference=LAUGIER C, PAROMTCHIK I E, PERROLLAZ M, et al. Probabilistic Analysis of Dynamic Scenes and Collision Risks Assessment to Improve Driving Safety[J]. IEEE Intelligent Transportation Systems Magazine, 2011, 3(4): 4-19., articleTitle=Probabilistic Analysis of Dynamic Scenes and Collision Risks Assessment to Improve Driving Safety, refAbstract=null), Reference(id=1211006375837176664, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, doi=null, pmid=null, pmcid=null, year=2008, volume=9, issue=1, pageStart=137, pageEnd=147, url=null, language=null, rfNumber=[3], rfOrder=2, authorNames=EIDEHALL A, PETERSSON L, journalName=IEEE Transactions on Intelligent Transportation Systems, refType=null, unstructuredReference=EIDEHALL A, PETERSSON L. Statistical Threat Assessment for General Road Scenes Using Monte Carlo Sampling[J]. IEEE Transactions on Intelligent Transportation Systems, 2008, 9(1): 137-147., articleTitle=Statistical Threat Assessment for General Road Scenes Using Monte Carlo Sampling, refAbstract=null), Reference(id=1211006375904285529, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, doi=null, pmid=null, pmcid=null, year=2012, volume=13, issue=2, pageStart=724, pageEnd=736, url=null, language=null, rfNumber=[4], rfOrder=3, authorNames=AOUDE G S, DESARAJU V R, STEPHENS L H, journalName=IEEE Transactions on Intelligent Transportation Systems, refType=null, unstructuredReference=AOUDE G S, DESARAJU V R, STEPHENS L H, et al. Driver Behavior Classification at Intersections and Validation on Large Naturalistic Data Set[J]. IEEE Transactions on Intelligent Transportation Systems, 2012, 13(2): 724-736., articleTitle=Driver Behavior Classification at Intersections and Validation on Large Naturalistic Data Set, refAbstract=null), Reference(id=1211006375971394394, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, doi=null, pmid=null, pmcid=null, year=2018, volume=null, issue=null, pageStart=1179, pageEnd=1184, url=null, language=null, rfNumber=[5], rfOrder=4, authorNames=DEO N, TRIVEDI M M, journalName=Changshu, refType=null, unstructuredReference=DEO N, TRIVEDI M M. Multi-Modal Trajectory Prediction of Surrounding Vehicles with Maneuver Based LSTMs[C]// 2018 IEEE Intelligent Vehicles Symposium (IV). Changshu, China: IEEE, 2018: 1179-1184., articleTitle=Multi-Modal Trajectory Prediction of Surrounding Vehicles with Maneuver Based LSTMs[C]// 2018 IEEE Intelligent Vehicles Symposium (IV), refAbstract=null), Reference(id=1211006376042697563, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, doi=null, pmid=null, pmcid=null, year=2017, volume=null, issue=null, pageStart=353, pageEnd=359, url=null, language=null, rfNumber=[6], rfOrder=5, authorNames=ALTCHÉ F, DE LA FORTELLE A, journalName=2017 IEEE 20th International Conference on Intelligent Transportation Systems (ITSC), refType=null, unstructuredReference=ALTCHÉ F, DE LA FORTELLE A. An LSTM Network for Highway Trajectory Prediction[C]// 2017 IEEE 20th International Conference on Intelligent Transportation Systems (ITSC). Yokohama, Japan: IEEE, 2017: 353-359., articleTitle=An LSTM Network for Highway Trajectory Prediction, refAbstract=null), Reference(id=1211006376118195036, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, doi=null, pmid=null, pmcid=null, year=2018, volume=null, issue=null, pageStart=1468, pageEnd=1476, url=null, language=null, rfNumber=[7], rfOrder=6, authorNames=DEO N, TRIVEDI M M, journalName=Salt Lake City, UT, refType=null, unstructuredReference=DEO N, TRIVEDI M M. Convolutional Social Pooling for Vehicle Trajectory Prediction[C]// 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW). Salt Lake City, UT, USA: IEEE, 2018: 1468-1476., articleTitle=Convolutional Social Pooling for Vehicle Trajectory Prediction[C]// 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW), refAbstract=null), Reference(id=1211006376189498205, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, doi=null, pmid=null, pmcid=null, year=2022, volume=16, issue=12, pageStart=1682, pageEnd=1696, url=null, language=null, rfNumber=[8], rfOrder=7, authorNames=CHEN L, ZHOU Q Y, CAI Y F, journalName=IET Intelligent Transport Systems, refType=null, unstructuredReference=CHEN L, ZHOU Q Y, CAI Y F, et al. CAE-GAN: A Hybrid Model for Vehicle Trajectory Prediction[J]. IET Intelligent Transport Systems, 2022, 16(12): 1682-1696., articleTitle=CAE-GAN: A Hybrid Model for Vehicle Trajectory Prediction, refAbstract=null), Reference(id=1211006376260801374, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=1, pageEnd=10, url=null, language=null, rfNumber=[9], rfOrder=8, authorNames=黄玲, 崔躜, 游峰, journalName=吉林大学学报(工学版), refType=null, unstructuredReference=黄玲, 崔躜, 游峰, 等. 适用于多车交互场景的车辆轨迹预测模型[J/OL]. 吉林大学学报(工学版): 1-10( 2022-12-07)[2023-08-21]. https://doi.org/10.13229/j.cnki.jdxbgxb20220728., articleTitle=适用于多车交互场景的车辆轨迹预测模型, refAbstract=null), Reference(id=1211006376327910239, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=1, pageEnd=10, url=null, language=null, rfNumber=[9], rfOrder=9, authorNames=HUANG L, CUI Z, YOU F, journalName=Journal of Jilin University (Engineering and Technology Edition), refType=null, unstructuredReference=HUANG L, CUI Z, YOU F, et al. Vehicle Trajectory Prediction Model for Multi-Vehicle Interaction Scenario[J/OL]. Journal of Jilin University (Engineering and Technology Edition): 1-10( 2022-12-07)[2023-08-21]. https://doi.org/10.13229/j.cnki.jdxbgxb20220728., articleTitle=Vehicle Trajectory Prediction Model for Multi-Vehicle Interaction Scenario, refAbstract=null), Reference(id=1211006376390824800, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, doi=null, pmid=null, pmcid=null, year=2020, volume=null, issue=null, pageStart=541, pageEnd=556, url=null, language=null, rfNumber=[10], rfOrder=10, authorNames=LIANG M, YANG B, HU R, journalName=European Conference on Computer Vision. Glasgow, Scotland, refType=null, unstructuredReference=LIANG M, YANG B, HU R, et al. Learning Lane Graph Representations for Motion Forecasting[C]// European Conference on Computer Vision. Glasgow, Scotland, UK: Springer, Cham, 2020: 541-556., articleTitle=Learning Lane Graph Representations for Motion Forecasting, refAbstract=null), Reference(id=1211006376466322273, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=https://arxiv.org/abs/1907.07792., language=null, rfNumber=[11], rfOrder=11, authorNames=LI X, YING X W, CHUAH M C, journalName=null, refType=null, unstructuredReference=LI X, YING X W, CHUAH M C. Grip++: Enhanced Graph-Based Interaction-Aware Trajectory Prediction for Autonomous Driving[EB/OL]. ( 2020-05-19)[2023-08-22]., articleTitle=Enhanced Graph-Based Interaction-Aware Trajectory Prediction for Autonomous Driving, refAbstract=null), Reference(id=1211006376541819746, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, doi=null, pmid=null, pmcid=null, year=2022, volume=null, issue=null, pageStart=9107, pageEnd=9114, url=null, language=null, rfNumber=[12], rfOrder=12, authorNames=GILLES T, SABATINI S, TSISHKOU D, journalName=null, refType=null, unstructuredReference=GILLES T, SABATINI S, TSISHKOU D, et al. GOHOME:Graph-Oriented Heatmap Output for Future Motion Estimation[C]// 2022 International Conference on Robotics and Automation (ICRA). Philadelphia, PA, USA: IEEE, 2022: 9107-9114., articleTitle=GOHOME:Graph-Oriented Heatmap Output for Future Motion Estimation[C], refAbstract=null), Reference(id=1211006376604734307, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, doi=null, pmid=null, pmcid=null, year=2023, volume=null, issue=7, pageStart=24, pageEnd=30, url=null, language=null, rfNumber=[13], rfOrder=13, authorNames=赵树恩, 苏天彬, 赵东宇, journalName=汽车技术, refType=null, unstructuredReference=赵树恩, 苏天彬, 赵东宇. 基于图神经网络的交互车辆驾驶意图识别及轨迹预测[J]. 汽车技术, 2023(7): 24-30., articleTitle=基于图神经网络的交互车辆驾驶意图识别及轨迹预测, refAbstract=null), Reference(id=1211006376663454564, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, doi=null, pmid=null, pmcid=null, year=2023, volume=null, issue=7, pageStart=24, pageEnd=30, url=null, language=null, rfNumber=[13], rfOrder=14, authorNames=ZHAO S E, SU T B, ZHAO D Y, journalName=Automobile Technology, refType=null, unstructuredReference=ZHAO S E, SU T B, ZHAO D Y. Interactive Vehicle Driving Intention Recognition and Trajectory Prediction Based on Graph Neural Network[J]. Automobile Technology, 2023(7): 24-30, articleTitle=Interactive Vehicle Driving Intention Recognition and Trajectory Prediction Based on Graph Neural Network, refAbstract=null), Reference(id=1211006376730563429, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, doi=null, pmid=null, pmcid=null, year=2022, volume=null, issue=null, pageStart=8823, pageEnd=8833, url=null, language=null, rfNumber=[14], rfOrder=15, authorNames=ZHOU Z K, YE L Y, WANG J P, journalName=2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition. New Orleans, LA, refType=null, unstructuredReference=ZHOU Z K, YE L Y, WANG J P, et al. HiVT: Hierarchical Vector Transformer for Multi-Agent Motion Prediction[C]// 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition. New Orleans, LA, USA: IEEE, 2022: 8823-8833., articleTitle=HiVT: Hierarchical Vector Transformer for Multi-Agent Motion Prediction, refAbstract=null), Reference(id=1211006376789283686, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, doi=null, pmid=null, pmcid=null, year=2023, volume=36, issue=1, pageStart=226, pageEnd=239, url=null, language=null, rfNumber=[15], rfOrder=16, authorNames=李文礼, 韩迪, 石晓辉, journalName=中国公路学报, refType=null, unstructuredReference=李文礼, 韩迪, 石晓辉, 等. 基于时-空注意力机制的车辆轨迹预测[J]. 中国公路学报, 2023, 36(1): 226-239., articleTitle=基于时-空注意力机制的车辆轨迹预测, refAbstract=null), Reference(id=1211006376860586855, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, doi=null, pmid=null, pmcid=null, year=2023, volume=36, issue=1, pageStart=226, pageEnd=239, url=null, language=null, rfNumber=[15], rfOrder=17, authorNames=LI W L, HAN D, SHI X H, journalName=China Journal of Highway and Transport, refType=null, unstructuredReference=LI W L, HAN D, SHI X H, et al. Vehicle Trajectory Prediction Based on Spatial-Temporal Attention Mechanism[J]. China Journal of Highway and Transport, 2023, 36(1): 226-239., articleTitle=Vehicle Trajectory Prediction Based on Spatial-Temporal Attention Mechanism, refAbstract=null), Reference(id=1211006376919307112, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, doi=null, pmid=null, pmcid=null, year=2023, volume=44, issue=7, pageStart=2162, pageEnd=2170, url=null, language=null, rfNumber=[16], rfOrder=18, authorNames=连静, 丁荣琪, 李琳辉, journalName=兵工学报, refType=null, unstructuredReference=连静, 丁荣琪, 李琳辉, 等. 基于图模型和注意力机制的车辆轨迹预测方法[J]. 兵工学报, 2023, 44(7): 2162-2170., articleTitle=基于图模型和注意力机制的车辆轨迹预测方法, refAbstract=null), Reference(id=1211006376978027369, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, doi=null, pmid=null, pmcid=null, year=2023, volume=44, issue=7, pageStart=2162, pageEnd=2170, url=null, language=null, rfNumber=[16], rfOrder=19, authorNames=LIAN J, DING R Q, LI L H, journalName=Acta Armamentarii, refType=null, unstructuredReference=LIAN J, DING R Q, LI L H, et al. Vehicle Trajectory Prediction Method Based on Graph Models and Attention Mechanism[J]. Acta Armamentarii, 2023, 44(7): 2162-2170., articleTitle=Vehicle Trajectory Prediction Method Based on Graph Models and Attention Mechanism, refAbstract=null), Reference(id=1211006377040941930, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, doi=null, pmid=null, pmcid=null, year=2019, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=[17], rfOrder=20, authorNames=KOSARAJU V, SADEGHIAN A, MARTÍN-MARTÍN R, journalName=Advances in Neural Information Processing Systems 2019. Vancouver, British Columbia, refType=null, unstructuredReference=KOSARAJU V, SADEGHIAN A, MARTÍN-MARTÍN R, et al. Social-BiGAT: Multimodal Trajectory Forecasting Using Bicycle-GAN and Graph Attention Networks[C]// Advances in Neural Information Processing Systems 2019. Vancouver, British Columbia, Canada: IEEE, 2019., articleTitle=Social-BiGAT: Multimodal Trajectory Forecasting Using Bicycle-GAN and Graph Attention Networks, refAbstract=null), Reference(id=1211006377103856491, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, doi=null, pmid=null, pmcid=null, year=2021, volume=48, issue=增刊1, pageStart=334, pageEnd=341, url=null, language=null, rfNumber=[18], rfOrder=21, authorNames=曾伟良, 陈漪皓, 姚若愚, journalName=计算机科学, refType=null, unstructuredReference=曾伟良, 陈漪皓, 姚若愚, 等. 时空图注意力网络在交叉口车辆轨迹预测的应用[J]. 计算机科学, 2021, 48(增刊1): 334-341., articleTitle=时空图注意力网络在交叉口车辆轨迹预测的应用, refAbstract=null), Reference(id=1211006377166771052, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, doi=null, pmid=null, pmcid=null, year=2021, volume=48, issue=S1, pageStart=334, pageEnd=341, url=null, language=null, rfNumber=[18], rfOrder=22, authorNames=ZENG W L, CHEN Y H, YAO R Y, journalName=Computer Science, refType=null, unstructuredReference=ZENG W L, CHEN Y H, YAO R Y, et al. Application of Spatial-Temporal Graph Attention Networks in Trajectory Prediction for Vehicles at Intersections[J]. Computer Science, 2021, 48(S1): 334-341., articleTitle=Application of Spatial-Temporal Graph Attention Networks in Trajectory Prediction for Vehicles at Intersections, refAbstract=null), Reference(id=1211006377233879917, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=https://arxiv.org/abs/2105.14491, language=null, rfNumber=[19], rfOrder=23, authorNames=BRODY S, ALON U, YAHAV E, journalName=null, refType=null, unstructuredReference=BRODY S, ALON U, YAHAV E. How Attentive Are Graph Attention Networks?[EB/OL]. ( 2022-01-31)[2023-08-22]., articleTitle=How Attentive Are Graph Attention Networks, refAbstract=null), Reference(id=1211006377296794478, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=https://arxiv.org/abs/1909.01315v2., language=null, rfNumber=[20], rfOrder=24, authorNames=WANG M J, ZHENG D, YE Z H, journalName=null, refType=null, unstructuredReference=WANG M J, ZHENG D, YE Z H, et al. Deep Graph Library: A Graph-Centric, Highly-Performant Package for Graph Neural Networks[EB/OL]. ( 2020-08-25)[2023-08-22]., articleTitle=A Graph-Centric, Highly-Performant Package for Graph Neural Networks, refAbstract=null), Reference(id=1211006378483782511, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, doi=null, pmid=null, pmcid=null, year=2007, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=[21], rfOrder=25, authorNames=COLYAR J, HALKIAS J, journalName=Washington, DC, refType=null, unstructuredReference=COLYAR J, HALKIAS J. US Highway 101 Dataset: FHWA-HRT-07-030[R]. Washington, DC: U.S. Department of Transportation (USDOT), 2007., articleTitle=US Highway 101 Dataset: FHWA-HRT-07-030, refAbstract=null), Reference(id=1211006378576057200, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, doi=null, pmid=null, pmcid=null, year=2007, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=[22], rfOrder=26, authorNames=COLYAR J, HALKIAS J, journalName=Washington, DC, refType=null, unstructuredReference=COLYAR J, HALKIAS J. US Highway I-80 Dataset: FHWA-HRT-07-030[R]. Washington, DC: U.S. Department of Transportation (USDOT), 2007., articleTitle=US Highway I-80 Dataset: FHWA-HRT-07-030, refAbstract=null), Reference(id=1211006378634777457, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, doi=null, pmid=null, pmcid=null, year=2016, volume=null, issue=null, pageStart=961, pageEnd=971, url=null, language=null, rfNumber=[23], rfOrder=27, authorNames=ALAHI A, GOEL K, RAMANATHAN V, journalName=2016 IEEE Conference on Computer Vision and Pattern Recognition. Las Vegas, NV, refType=null, unstructuredReference=ALAHI A, GOEL K, RAMANATHAN V, et al. Social LSTM: Human Trajectory Prediction in Crowded Spaces[C]// 2016 IEEE Conference on Computer Vision and Pattern Recognition. Las Vegas, NV, USA: IEEE, 2016: 961-971., articleTitle=Social LSTM: Human Trajectory Prediction in Crowded Spaces, refAbstract=null), Reference(id=1211006378693497714, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, doi=null, pmid=null, pmcid=null, year=2022, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=[24], rfOrder=28, authorNames=石振周, journalName=基于车道线信息的多目标轨迹预测, refType=null, unstructuredReference=石振周. 基于车道线信息的多目标轨迹预测[D]. 大连: 大连理工大学, 2022., articleTitle=null, refAbstract=null), Reference(id=1211006378752217971, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, doi=null, pmid=null, pmcid=null, year=2022, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=[24], rfOrder=29, authorNames=SHI Z Z, journalName=Multi-Objective Trajectory Prediction Based on Lane Information, refType=null, unstructuredReference=SHI Z Z. Multi-Objective Trajectory Prediction Based on Lane Information[D]. Dalian: Dalian University of Technology, 2022., articleTitle=null, refAbstract=null)], funds=[Fund(id=1211006375396774738, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, awardId=2021YFB1600501, language=CN, fundingSource=*国家重点研发计划项目(2021YFB1600501), fundOrder=null, country=null), Fund(id=1211006375455494995, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, awardId=61906038, language=CN, fundingSource=国家自然科学基金项目(61906038), fundOrder=null, country=null), Fund(id=1211006375526798164, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, awardId=null, language=CN, fundingSource=东南大学“至善青年学者”支持计划项目, fundOrder=null, country=null), Fund(id=1211006375581324117, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, awardId=2242021R41184, language=CN, fundingSource=中央高校基本科研业务费专项资金项目(2242021R41184), fundOrder=null, country=null)], companyList=[AuthorCompany(id=1211006372066497323, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, xref=null, ext=[AuthorCompanyExt(id=1211006372074885932, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, companyId=1211006372066497323, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=Southeast University, Nanjing 210096), AuthorCompanyExt(id=1211006372083274541, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, companyId=1211006372066497323, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=东南大学,南京 210096)])], figs=[ArticleFig(id=1211006374541136710, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, language=EN, label=null, caption=null, figureFileSmall=dok3WAzezlDzyu3Dsb7UQA==, figureFileBig=wLhm0tz6YvMrXAvZmKbheQ==, tableContent=null), ArticleFig(id=1211006374595662663, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, language=CN, label=图1, caption=ED-DGAT模型结构, figureFileSmall=dok3WAzezlDzyu3Dsb7UQA==, figureFileBig=wLhm0tz6YvMrXAvZmKbheQ==, tableContent=null), ArticleFig(id=1211006374671160136, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, language=EN, label=null, caption=null, figureFileSmall=izEwLh3iZ+emh3lAJM7/RA==, figureFileBig=RJ82xuN1EFbnbK7bmbItuw==, tableContent=null), ArticleFig(id=1211006374729880393, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, language=CN, label=图2, caption=ED-DGAT模型预测结果可视化, figureFileSmall=izEwLh3iZ+emh3lAJM7/RA==, figureFileBig=RJ82xuN1EFbnbK7bmbItuw==, tableContent=null), ArticleFig(id=1211006374788600650, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, language=EN, label=null, caption=null, figureFileSmall=t1fnX+7HCVvyMmPAutWcCw==, figureFileBig=d/DD9sDxK6g7ux+jCv1ncg==, tableContent=null), ArticleFig(id=1211006374843126603, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, language=CN, label=图3, caption=不同Dclose下模型性能比较, figureFileSmall=t1fnX+7HCVvyMmPAutWcCw==, figureFileBig=d/DD9sDxK6g7ux+jCv1ncg==, tableContent=null), ArticleFig(id=1211006374897652556, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, language=EN, label=null, caption=null, figureFileSmall=null, figureFileBig=null, tableContent=
设备名称 设备型号
中央处理器(CPU) I9-10900k
图形处理器(GPU) 4×GeForce GTX3080
系统 Ubuntu 18.04
内存 4×16 GB
统一计算架构(CUDA)版本 11.4
), ArticleFig(id=1211006374956372813, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, language=CN, label=表1, caption=

硬件试验平台

, figureFileSmall=null, figureFileBig=null, tableContent=
设备名称 设备型号
中央处理器(CPU) I9-10900k
图形处理器(GPU) 4×GeForce GTX3080
系统 Ubuntu 18.04
内存 4×16 GB
统一计算架构(CUDA)版本 11.4
), ArticleFig(id=1211006375023481678, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, language=EN, label=null, caption=null, figureFileSmall=null, figureFileBig=null, tableContent=
模型名称 ermse/m
预测时长1 s 预测时长2 s 预测时长3 s 预测时长4 s 预测时长5 s
V-LSTM 0.65 1.68 0.61 2.94 6.37
S-LSTM 0.65 1.31 2.16 3.25 4.55
CS-LSTM 0.61 1.27 2.09 3.10 4.37
ED-GCN 0.58 1.35 2.28 3.41 4.78
ED-GAT 0.61 1.33 2.16 3.10 4.21
E-DGAT 0.57 1.20 1.92 2.77 3.80
ED-DGAT 0.53 1.12 1.76 2.46 3.27
), ArticleFig(id=1211006375103173455, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, language=CN, label=表2, caption=

ED-DGAT在NGSIM数据集上试验结果

, figureFileSmall=null, figureFileBig=null, tableContent=
模型名称 ermse/m
预测时长1 s 预测时长2 s 预测时长3 s 预测时长4 s 预测时长5 s
V-LSTM 0.65 1.68 0.61 2.94 6.37
S-LSTM 0.65 1.31 2.16 3.25 4.55
CS-LSTM 0.61 1.27 2.09 3.10 4.37
ED-GCN 0.58 1.35 2.28 3.41 4.78
ED-GAT 0.61 1.33 2.16 3.10 4.21
E-DGAT 0.57 1.20 1.92 2.77 3.80
ED-DGAT 0.53 1.12 1.76 2.46 3.27
), ArticleFig(id=1211006375178670928, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, language=EN, label=null, caption=null, figureFileSmall=null, figureFileBig=null, tableContent=
模型 推理时长
V-LSTM 22.23
S-LSTM 27.12
CS-LSTM 27.80
ED-DGAT 10.64
), ArticleFig(id=1211006375245779793, tenantId=1146029695717560320, journalId=1189621681917173762, articleId=1210314383859782475, language=CN, label=表3, caption=

模型推理时长 ms

, figureFileSmall=null, figureFileBig=null, tableContent=
模型 推理时长
V-LSTM 22.23
S-LSTM 27.12
CS-LSTM 27.80
ED-DGAT 10.64
)], attaches=null, journal=Journal(id=1149693407745847311, delFlag=0, nameCn=汽车技术, nameEn=Automobile Technology, nameHistory1=null, nameHistory2=null, issn=1000-3703, eissn=null, cn=22-1113/U, coden=null, periodic=0, language=CN, oaType=null, ccby=null, superviseOffice=null, ownerOffice=null, pubOffice=null, editorOffice=null, officeType=null, aims=null, clcCode=null, officeProv=null, officeCity=null, officeAddr=null, officeZip=null, officeEmail=null, officePhone=null, editDirector=null, officeDirector=null, officeDirectorPhone=null, officeStaffNum=null, officeEmpNum=null, coverPicUrl=rYFtDx/CU9+iX8QTM0ckbw==, journalPrice=null, startedYear=null, abbrevIsoEn=null, journalRemark=null, publicationField=null, createdTime=1752037868679, updatedTime=1761735668047, createdBy=18614031015, updatedBy=13701087609, firstLetterCn=A, firstLetterEn=A, subjectCode=Engineering, subjectName=Engineering, subjectCodeEn=Engineering, subjectNameEn=null, picCn=rYFtDx/CU9+iX8QTM0ckbw==, picEn=oFT2NmUwKPUjZ27C1+d9pw==, jcr=null, cjcr=null, exts=[JournalExt(id=1190368926564450443, language=CN, name=汽车技术, nameHistory1=null, nameHistory2=null, managedBy=, sponsoredBy=, publishedBy=, editorOffice=, officeProv=null, officeCity=null, officeAddr=, officeZip=, editDirector=, officeDirector=null, officePhone=null, coverPicUrl=null, journalRemark=, submitArticleUrl=null, websiteUrl=, createdTime=1761735668078, updatedTime=1761735668078, createdBy=13701087609, updatedBy=13701087609, submissionGuidelinesUrl=, submissionAuthorUrl=https://qcjs.cbpt.cnki.net/index.aspx?t=1, submissionEditorUrl=https://qcjs.cbpt.cnki.net/index.aspx?t=3, submissionReviewUrl=https://qcjs.cbpt.cnki.net/index.aspx?t=2, submissionCeEditorUrl=, submissionAeEditorUrl=, option={"copyright":""}), JournalExt(id=1190368926618976396, language=EN, name=Automobile Technology, nameHistory1=null, nameHistory2=null, managedBy=, sponsoredBy=, publishedBy=, editorOffice=, officeProv=null, officeCity=null, officeAddr=, officeZip=, editDirector=, officeDirector=null, officePhone=null, coverPicUrl=null, journalRemark=, submitArticleUrl=null, websiteUrl=, createdTime=1761735668091, updatedTime=1761735668091, createdBy=13701087609, updatedBy=13701087609, submissionGuidelinesUrl=, submissionAuthorUrl=https://qcjs.cbpt.cnki.net/index.aspx?t=1, submissionEditorUrl=https://qcjs.cbpt.cnki.net/index.aspx?t=3, submissionReviewUrl=https://qcjs.cbpt.cnki.net/index.aspx?t=2, submissionCeEditorUrl=, submissionAeEditorUrl=, option={"copyright":""})], databaseList=null, tenantJournalId=1189621681917173762, websiteList=[Website(id=1189624193747526544, webName=null, webTitle=null, webDomain=null, webCopyrigh=null, webIpcNo=null, seoTitle=null, seoKeywords=null, seoDescription=null, tenantJournalId=null, journalId=1189621681917173762, journalNameCn=null, journalNameEn=null, grayFlag=null, tenantId=1146029695717560320, platformId=null, journalGroupId=null, journalGroupNameCn=null, journalGroupNameEn=null, type=1, domain=https://castjournals.cast.org.cn/joweb/qcjs/CN, language=CN, createTime=1761558109939, createBy=18614031015, updateTime=1761558140534, updateBy=18614031015, name=汽车技术-中, tplId=1146099689490845704, title=汽车技术, delFlag=0, indexPage=/home, props=[WebsiteProps(id=1189625424704451180, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1189624193747526544, code=articleTextType, value=kx, createTime=1761558403421, updateTime=1761558403421, creator=18614031015, updator=18614031015), WebsiteProps(id=1189625424675091049, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1189624193747526544, code=banner, value=null, createTime=1761558403414, updateTime=1761558403414, creator=18614031015, updator=18614031015), WebsiteProps(id=1189625424733811311, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1189624193747526544, code=grayFlag, value=0, createTime=1761558403428, updateTime=1761558403428, creator=18614031015, updator=18614031015), WebsiteProps(id=1189625424658313832, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1189624193747526544, code=logo, value=https://castjournals.cast.org.cn/joweb/qcjs/CN/file/pic?fileId=7En9rzX2QCa/1J8NnKt/Fg==, createTime=1761558403410, updateTime=1761558403410, creator=18614031015, updator=18614031015), WebsiteProps(id=1189625424746394225, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1189624193747526544, code=minRunFlag, value=0, createTime=1761558403431, updateTime=1761558403431, creator=18614031015, updator=18614031015), WebsiteProps(id=1189625424691868267, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1189624193747526544, code=picServerUrl, value=https://castjournals.cast.org.cn/joweb/qcjs/CN/file/pic, createTime=1761558403418, updateTime=1761558403418, creator=18614031015, updator=18614031015), WebsiteProps(id=1189625424742199920, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1189624193747526544, code=silenceFlag, value=0, createTime=1761558403430, updateTime=1761558403430, creator=18614031015, updator=18614031015), WebsiteProps(id=1189625424683479658, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1189624193747526544, code=staticResourcePath, value=https://castjournals.cast.org.cn/joweb/cast_kjdb_cn_619/, createTime=1761558403416, updateTime=1761558403416, creator=18614031015, updator=18614031015), WebsiteProps(id=1189625424712839789, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1189624193747526544, code=themeColor, value=null, createTime=1761558403423, updateTime=1761558403423, creator=18614031015, updator=18614031015), WebsiteProps(id=1189625424725422702, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1189624193747526544, code=themeStyle, value=null, createTime=1761558403426, updateTime=1761558403426, creator=18614031015, updator=18614031015)]), Website(id=1189624193869161363, webName=null, webTitle=null, webDomain=null, webCopyrigh=null, webIpcNo=null, seoTitle=null, seoKeywords=null, seoDescription=null, tenantJournalId=null, journalId=1189621681917173762, journalNameCn=null, journalNameEn=null, grayFlag=null, tenantId=1146029695717560320, platformId=null, journalGroupId=null, journalGroupNameCn=null, journalGroupNameEn=null, type=1, domain=https://castjournals.cast.org.cn/joweb/qcjs/EN, language=EN, createTime=1761558109967, createBy=18614031015, updateTime=1761558340679, updateBy=18614031015, name=汽车技术-英文, tplId=1146101810881728533, title=Automobile Technology, delFlag=0, indexPage=/home, props=[WebsiteProps(id=1189625550722311064, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1189624193869161363, code=articleTextType, value=kx, createTime=1761558433466, updateTime=1761558433466, creator=18614031015, updator=18614031015), WebsiteProps(id=1189625550688756629, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1189624193869161363, code=banner, value=null, createTime=1761558433458, updateTime=1761558433458, creator=18614031015, updator=18614031015), WebsiteProps(id=1189625550739088283, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1189624193869161363, code=grayFlag, value=0, createTime=1761558433470, updateTime=1761558433470, creator=18614031015, updator=18614031015), WebsiteProps(id=1189625550676173716, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1189624193869161363, code=logo, value=https://castjournals.cast.org.cn/joweb/qcjs/EN/file/pic?fileId=7En9rzX2QCa/1J8NnKt/Fg==, createTime=1761558433455, updateTime=1761558433455, creator=18614031015, updator=18614031015), WebsiteProps(id=1189625550751671197, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1189624193869161363, code=minRunFlag, value=0, createTime=1761558433473, updateTime=1761558433473, creator=18614031015, updator=18614031015), WebsiteProps(id=1189625550713922455, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1189624193869161363, code=picServerUrl, value=https://castjournals.cast.org.cn/joweb/qcjs/EN/file/pic, createTime=1761558433464, updateTime=1761558433464, creator=18614031015, updator=18614031015), WebsiteProps(id=1189625550743282588, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1189624193869161363, code=silenceFlag, value=0, createTime=1761558433471, updateTime=1761558433471, creator=18614031015, updator=18614031015), WebsiteProps(id=1189625550705533846, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1189624193869161363, code=staticResourcePath, value=https://castjournals.cast.org.cn/joweb/cast_kjdb_en_623/, createTime=1761558433462, updateTime=1761558433462, creator=18614031015, updator=18614031015), WebsiteProps(id=1189625550726505369, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1189624193869161363, code=themeColor, value=null, createTime=1761558433467, updateTime=1761558433467, creator=18614031015, updator=18614031015), WebsiteProps(id=1189625550734893978, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1189624193869161363, code=themeStyle, value=null, createTime=1761558433469, updateTime=1761558433469, creator=18614031015, updator=18614031015)])], journalTitle=汽车技术, weixinUrl=null, journalUrl=null, iacademicId=null, status=1, seqNo=null, journalTitleEn=Automobile Technology, journalPhotoCn=rYFtDx/CU9+iX8QTM0ckbw==, journalPhotoEn=oFT2NmUwKPUjZ27C1+d9pw==, journalFirstLetter=A, journalRecommend=null, journalNew=null, journalCollection=null, jcrJf=null, cjcrJf=null, jcrJfStr=null, cjcrJfStr=null, submissionFirstDecision=null, sciSubjectClassification=null, casSubjectClassification=null, citeScore=null, totalCitationFrequency=null, icpCode=null, psCode=null, advertisingLicenseCode=null, copyrightInformation=null, country=null, option=, provinceCode=null, provinceName=null, collectFlag=false), detailUrlCn=https://castjournals.cast.org.cn/joweb/qcjs/CN/10.19620/j.cnki.1000-3703.20230582, detailUrlEn=https://castjournals.cast.org.cn/joweb/qcjs/EN/10.19620/j.cnki.1000-3703.20230582, pdfUrlCn=https://castjournals.cast.org.cn/joweb/qcjs/CN/PDF/10.19620/j.cnki.1000-3703.20230582, pdfUrlEn=https://castjournals.cast.org.cn/joweb/qcjs/EN/PDF/10.19620/j.cnki.1000-3703.20230582, aliStartDate=null, aliEndDate=null, collectionFlag=false, citedCount=null, citedUrl=null, reference=null)
收藏切换
基于动态图注意力的车辆轨迹预测研究*
收藏切换
PDF下载
陈晓伟 , 李煊鹏 , 张为公
汽车技术 | 2024,(3): 24-30
收起
收藏切换
汽车技术 | 2024, (3): 24-30
基于动态图注意力的车辆轨迹预测研究*
全屏
陈晓伟, 李煊鹏, 张为公
作者信息
  • 东南大学,南京 210096
Research on Vehicle Trajectory Prediction Based on Dynamic Graph Attention
Xiaowei Chen, Xuanpeng Li, Weigong Zhang
Affiliations
  • Southeast University, Nanjing 210096
出版时间: 2024-03-24 doi: 10.19620/j.cnki.1000-3703.20230582
文章导航
收藏切换

针对目前轨迹预测研究中交互建模方法使用的图注意力网络(GAT)为静态注意力,无法有效捕捉复杂道路场景中车辆间交互的问题,提出了一种基于编码器-解码器架构的动态图注意力网络(ED-DGAT)预测高速公路环境中运动车辆的未来轨迹。编码模块使用动态图注意力机制学习场景中车辆间的空间交互,采用状态简化动态图注意力网络建模解码阶段车辆运动的相互依赖,最后使用NGSIM数据集评估所提出的模型,并与长短时记忆(LSTM)、联合社交池化与长短时记忆(S-LSTM)、联合卷积社交池化与长短时记忆(CS-LSTM)算法模型进行对比分析,结果表明,预测轨迹的均方根误差(RMSE)降低了25%,且模型的推理速度为CS-LSTM模型的2.61倍。

轨迹预测  /  注意力机制  /  图神经网络  /  多目标交互

In current research on vehicle trajectory prediction, the existing Graph Attention Network (GAT), which is based on a static attention mechanism, fails to effectively capture interactions between vehicles in complex road conditions. To address this issue, this paper proposed an Encoder-Decoder Dynamic Graph Attention Network (ED-DGAT) to predict future trajectories of highway vehicles. In this model, the encoding module incorporates a dynamic graph attention mechanism to learn spatial interactions among vehicles. Simultaneously, a simplified dynamic graph attention network is adopted to model the interdependencies of vehicle movements during the decoding phase. This paper evaluated the proposed algorithm using the NGSIM dataset and conducted comparative analysis with other models such as LSTM, Social-LSTM (S-LSTM), and CS-LSTM. The results show that the Root Mean Squared Error (RMSE) of predicted trajectory has been reduced by 25%, and the inference speed is 2.61 times of the CS-LSTM model.

Trajectory prediction  /  Attention mechanism  /  Graph neural networks  /  Multi-objective interaction
陈晓伟, 李煊鹏, 张为公. 基于动态图注意力的车辆轨迹预测研究*. 汽车技术, 2024 , (3) : 24 -30 . DOI: 10.19620/j.cnki.1000-3703.20230582
Xiaowei Chen, Xuanpeng Li, Weigong Zhang. Research on Vehicle Trajectory Prediction Based on Dynamic Graph Attention[J]. Automobile Technology, 2024 , (3) : 24 -30 . DOI: 10.19620/j.cnki.1000-3703.20230582
为实现安全高效行驶,自动驾驶汽车需自主且合理地估计周围目标的运动轨迹。然而,车辆的行为决策不仅与驾驶员的驾驶风格、道路环境相关,还受到周围车辆间交互行为的影响,故预测复杂道路场景中车辆的运动轨迹极具挑战性。
传统方法基于运动物理学[1]或结合概率论[2-4]估计车辆的运动轨迹。随着人工智能技术的发展,许多研究人员开始采用基于循环神经网络(Recurrent Neural Network,RNN)及其变体的编码器-解码器架构进行时序数据的预测研究[5-6]。然而,RNN无法捕捉运动车辆间的空间相关性。
文献[7]~文献[9]引入了一种名为“社交池化”的机制,该方法将道路中相邻车辆的运动特征聚合到统一的空间网格,并利用池化和卷积等方法融合车辆的运动特征。文献[10]~文献[13]基于图神经网络(Graph Neural Network,GNN)建模车辆间的交互关系,其中图的节点表示运动车辆,图的边表示车辆间的空间交互,并利用图卷积网络(Graph Convolutional Network,GCN)聚合相邻车辆的特征。文献[14]~文献[16]基于注意力机制加权聚合邻近车辆的相互影响,重点关注对目标车辆具有更高影响力的邻近车辆。与“社交池化”方法和基于图的方法相比,基于注意力机制的方法可以更好地解释运动车辆间的空间交互。
此外,上述所有方法只考虑历史轨迹编码阶段车辆间的交互,而在预测阶段建模邻近车辆的影响至关重要。
文献[17]、文献[18]利用基于注意力机制的GNN变体——图注意力网络(Graph Attention Network,GAT)建模道路目标间的交互。Brody等[19]从概念上证明了GAT计算的是静态注意力,并提出了基于动态注意力机制的GATv2(本文称为动态图注意力网络)。与GAT相比,动态图注意力网络(Dynamic Graph Attention Network,DGAT)具备更强的图表示能力。受此启发,本文提出一种基于编码器-解码器架构的动态图注意力网络(Encoder-Decoder Dynamic Graph Attention Network,ED-DGAT)预测车辆轨迹,模型基于RNN的变体-循环门控单元(Gate Recurrent Unit,GRU)编码单一车辆运动的时间相关性,设计基于DGAT的交互建模模块,利用DGAT聚合GRU的隐藏状态并以此编码相邻车辆的历史运动状态,在GRU解码阶段的每一时间帧使用状态简化DGAT捕获车辆间的空间交互。
本文的目标是根据高速公路场景中目标车辆与邻近车辆的历史轨迹信息,预测得到车辆的未来轨迹。车辆的历史轨迹可以使用一系列离散的点描述,本文所提出算法模型的输入为:
X = x T o b s - t h , x T o b s - t h + 1 , , x T o b s
式中:n为道路中车辆数量,Tobs为道路场景的观察时刻,th为历史轨迹时长,xt=(x0t,x1t,…,xnt)为t时刻道路所有车辆的位置信息,xit为车辆it时刻的坐标。
与之前的工作[5,7]只预测单一车辆的未来轨迹不同,本文选择一次性预测道路场景中所有车辆的未来轨迹:
Y = y 1 , y 2 , , y t f
式中:tf为预测时长,yt=(y0t,y1t,…,ynt)为未来时刻t道路中所有车辆的位置信息,yit为车辆it时刻的坐标。
相应地,使用 Y ^ = y ^ 1 , y ^ 2 , , y ^ t f表示模型的输出,其中, y ^ t为未来时刻t道路中所有车辆位置的预测值,使用 y ^ i t表示t时刻车辆i坐标的预测值。
本文所提出的算法模型ED-DGAT由历史轨迹编码模块、交互编码模块和预测解码模块组成,具体结构如图1所示。历史轨迹编码模块基于GRU编码单一车辆的历史轨迹,交互编码模块利用DGAT提取可变数量的邻近车辆之间的交互特征,预测解码模块结合单一车辆的历史轨迹编码与邻近车辆的交互编码,生成车辆的未来轨迹。
ED-DGAT模型使用双层GRU编码单一车辆的运动特征,所有车辆共享同一个GRU,GRU的隐藏维度为64。对于Tobs时刻,使用多层感知机(Multi-Layer Perceptron,MLP)编码最近th帧的车辆轨迹,然后按时间顺序输入GRU中,最后得到隐藏状态矩阵H

H=(h0,h1,…,hn)

式中:hi为车辆i的运动特征。
在道路场景中,车辆的运动受周围车辆的影响,这和人与人之间的社交网络相似,研究人员常使用图神经网络建模人与人之间的社交网络,受此启发,本文选择使用图神经网络的最新方法DGAT建模车辆间的交互。DGAT为不同的邻近车辆分配不同的权重,在此基础上加权聚合邻近车辆的特征,从而更加合理地考虑周围车辆对目标车辆的影响。
DGAT的计算过程基于图结构,因此图的结构将显著影响模型的计算效率和预测性能。考虑到车辆间的影响互不相同,本文使用有向图G={V,E}表示道路场景中的车辆关系,其中VE分别为图的节点集和边集。
定义1(有向图):对于图G={V,E},设图中节点数量为m,则节点集V={1,…,m},边集E={e0,0,e0,1,…,ei,j,…,em,m}⊂V×V表示V中任意2个节点之间的边,其中ei,j表示从节点i到节点j的边。若从节点i到节点j的边与从节点j到节点i的边不相等,则图G为一个有向图。
如果图只包含自连接,模型的性能理论上应与无交互建模的模型相似。而如果采用全连接的方式,将会导致冗余连接,随着图中节点的数量增加,冗余连接将呈指数增长,严重影响模型的计算效率。因此,本文所构建的图神经网络为邻居节点指向中心节点的星形结构。
节点集合V中的节点表示道路场景中的运动车辆,节点特征vi为车辆i的历史轨迹编码。此外,本文使用图的第1个节点表示被预测车辆,其余节点{1,2,…,m}表示邻近车辆,因此边集合表示为:

E={ej,0}(j=0,…,m)

式中:ej,0表示图中的节点j和节点0之间存在有向边连接,节点j是节点0的邻居。
道路场景中,车辆的距离靠近就会产生交互,因此本文采取距离阈值Dclose判断2个节点间是否存在边连接,如果节点所表示的车辆在Tobs时刻的距离小于阈值,则认定节点间存在边连接,设Dclose=50 m。图1中,圆圈内中心处的车辆认定为与圈内其他车辆存在空间交互。
DGAT基于自注意力机制加权聚合邻居节点的特征,并用于更新节点自身。DGAT的输入为历史轨迹的编码H,模型输出为 H ' = ( h 0 ' , h 1 ' , , h n ' )。图神经网络中节点对(节点i与节点j)间的注意力系数为:
α i , j = s o f t   m a x ( e ( h i , h j ) ) = e x p ( e ( h i , h j ) ) k N i e x p ( e ( h i , h k ) )
式中:αij为节点i对节点j的注意力系数;Ni为图神经网络中节点i的邻居节点集合;e为评分函数,用于计算每条边的初始分数;softmax函数用于归一化节点i所有邻居节点的初始注意力系数。
评分函数的定义为:
e ( h i , h j ) = a T σ W h i W h j
式中:||表示向量的拼接,W∈RF×F为应用于每个节点的共享线性变换的权重矩阵,FF′分别为DGAT输入、输出的维度,a∈R2F为单层前馈神经网络的权重矩阵,σ为非线性激活函数LeakyReLU。
得到节点对其邻居节点的注意力系数后,DGAT通过加权求和邻居节点的特征更新中心节点的特征:
h i ' = σ k N i a i j W h j
式(5)~式(7)展示了单层DGAT的计算方法。本文采用单层DGAT构建车辆间的交互,DGAT的输出 h i '为邻近车辆对车辆i的空间交互特征。此外,模型中使用了多头注意力机制提升DGAT的性能,多头注意力机制的数量为8。
预测解码模块的输入为单一车辆的动态特征与邻近车辆的交互特征。将历史轨迹编码模型的输出H与交互编码模块的输出H′输入到2个不同的MLP中,并将输出拼接,以此融合车辆的运动模式与空间交互特征。具体计算流程为:
h i ¯ = δ 1 ( h i )
g i = δ 2 ( h i ' )
d i = h i ¯ g i
式中:δ1δ2分别为编码HH′的3层MLP,使用ReLU激活函数,δ1的隐含层维度为128和64,δ2的隐含层维度为256和128; h i ¯gi分别为δ1δ2的编码输出,维度分别为32和64;di为车辆i的解码输入。
与历史编码器相似,本文采用双层GRU作为预测解码模块的骨干网络,同时采用状态简化的DGAT(使用D-DGAT表示解码模块所使用的DGAT)建模预测阶段车辆间的交互。与编码动态图注意力(Encoder Dynamic Graph Attention Network,E-DGAT)不同,解码动态图注意力(Decoder Dynamic Graph Attention Network,D-DGAT)不再计算目标节点对邻居节点的注意力系数,而是直接采用E-DGAT计算的注意力系数。D-DGAT的输入为t时刻目标车辆与邻近车辆的坐标。具体计算流程为:
γ i t - 1 = η 1 ( y ^ i t - 1 )
γ ˜ i t - 1 = σ k N i a i j W d γ i t - 1
g ˜ i t - 1 = δ 3 ( γ ˜ i t - 1 )
式中:η1为与历史轨迹编码模块中相同的单层MLP,使用ReLU激活函数,输出维度为32;γjt-1为邻近车辆j在(t-1)时刻位置的编码;Wd为线性变换函数的权重矩阵; γ ˜ i t - 1为车辆i在(t-1)时刻的空间交互特征;δ3为3层MLP,使用ReLU激活函数,隐含层维度为256和128; g ˜ i t - 1δ3的输出,维度为64。
解码GRU的隐藏状态初始化为零,输入包括编码GRU的隐藏状态、交互编码模块的输出、上一时间帧目标车辆的位置坐标、上一时间帧邻近车辆的交互编码。当前时刻解码模块的计算方法为:
m i t = Φ ( d i , y ^ i t - 1 , g ˜ i t - 1 )
y ^ i t = η 2 ( m i t )
式中:Ф为解码GRU,隐藏维度为128;mit为车辆i在时刻t的GRU隐藏状态;η2为单层MLP。
具体而言,第1个解码步骤中使用Tobs时刻的车辆坐标作为输入,当前时间步的输出用作下一时间步解码GRU的输入。对应的解码步骤循环tf次,得到所有车辆的未来轨迹。
本文使用Adam作为优化器训练网络,学习率为0.001,模型每次训练1个驾驶场景,即批大小(Batchsize)设置为1,共训练5代(Epoch)。
本文提出的模型使用预测得到的轨迹输出和轨迹真值计算损失。与文献[7]相似,本文使用均方误差作为损失函数:
L = 1 n i n Y ^ i - Y i 2
式中:Y为车辆i的真实未来轨迹, Y ^ i为模型预测结果。
本文所有试验均在同一服务器上完成,具体硬件配置如表1所示。所有试验均基于Python语言和深度学习框架Pytorch实现,其中图神经网络相关试验基于深度图谱库(Deep Graph Library,DGL)[20]开源框架实现。
本文选择在NGSIM数据集上评估所提出的算法模型。NGSIM数据集广泛用于车辆轨迹预测,分为us-101[21]和i-80[22]2个部分,分别在长度为640 m和503 m的高速公路中以10 Hz的频率采集的总时长为45 min的真实车辆轨迹。将数据集分为训练集、验证集和测试集,比例为7∶1∶2。从数据集中选取连续的8 s片段,前3 s用作观察的轨迹历史,后5 s作为被预测的真实轨迹。与Deo等[6-7,11]相同,对数据进行下采样,采样频率为5 Hz。
本文试验采用均方根误差(Root Mean Squared Error,RMSE)评估模型性能:
e r m s e = i n t 1,2 , , t f y ^ i t - y i t 2 n t f
将本文提出的ED-DGAT模型与对比模型进行对比验证。对比模型包括单一长短时记忆(Vanilla LSTM,V-LSTM)、联合社交池化与长短时记忆(Social LSTM,S-LSTM)[23]、联合卷积社交池化与长短时记忆(LSTM with Convolutional Social pooling,CS-LSTM)、E-DGAT、编码-解码图卷积(Encoder-Decoder Graph Convolutional Network,ED-GCN)、编码-解码图注意力(Encoder-Decoder Graph Attention Network,ED-GAT),表2提供了上述模型在NGSIM数据集上预测时长分别为1~5 s的RMSE指标。
表2可知,与CS-LSTM模型相比,预测的车辆轨迹在5 s时长上的RMSE评价指标降低了25%。比较使用GAT作为图聚合方式的ED-GAT模型与使用池化方案的S-LSTM模型、CS-LSTM模型和使用GCN作为图聚合方式的ED-GCN模型,ED-GAT模型的RMSE最低,验证了基于注意力机制的交互建模方法可以更好地理解运动车辆间的空间交互。使用DGAT作为图聚合方式的ED-DGAT模型的预测误差进一步下降,原因在于:道路场景中车辆间空间交互的复杂程度高,DGAT的图表示能力更强,更加适用于车辆轨迹预测领域。ED-DGAT模型与E-DGAT模型的试验结果表明,在预测阶段建模车辆之间的交互可以有效提升模型的预测性能。
为了更好地理解DGAT在提高车辆间空间交互建模方面的贡献,本文对同一道路场景下不同车辆ED-DGAT模型的预测结果进行可视化,如图2所示。由图2可知,在高速公路场景中,同一车道的后方车辆跟随前方车辆行驶,对于变道车辆而言,以避免碰撞为原则汇入目标车道。图2中,对于直行车辆,ED-DGAT模型能有效捕捉车辆间的跟随行为,对于变道车辆,ED-DGAT模型能有效识别车辆的变道意图。对于直行车辆和变道车辆,ED-DGAT模型都能捕获车辆轨迹的细节,得到更加贴近真实场景的预测结果。
本文将Dclose设置为不同值,探讨其对模型性能的影响,试验结果如图3所示。从图3中可以看出,相较于Dclose=0,Dclose>0时预测结果的5 s RMSE显著降低,原因在于:Dclose=0时目标车辆不考虑邻近车辆的影响,模型等同于使用单一GRU的编码器-解码器结构。上述结果表明,建模车辆间的交互能有效降低模型的预测误差。对比Dclose=100 m与Dclose=50 m模型的性能,Dclose=100 m时模型的RMSE较低,原因在于:考虑较大的Dclose阈值,模型会考虑更多邻近车辆对目标车辆的影响。然而,Dclose为100 m时RMSE只在Dclose为50 m的基础上略有降低,原因在于:当Dclose>50 m时,模型已经充分考虑了对目标车辆影响最大的邻近车辆,此时再增大Dclose,模型虽然会考虑更多的邻近车辆对目标车辆的影响,但这些车辆距离目标车辆较远,对目标车辆的影响远低于近距离车辆。此外,考虑更大的Dclose,DGAT的计算量将会大幅度增长。综合考虑,本文采用Dclose=50 m作为试验的基本设置。
衡量车辆轨迹预测模型的标准除预测轨迹的误差外,还应关注模型的计算效率。石振周[24]等验证了模型推理时间在40 ms内即可满足实时性要求。考虑到自动驾驶应用场景中,一般使用嵌入式平台运行模型,计算资源有限,本文在单块GPU上验证算法模型,模拟实际驾驶场景中的车辆轨迹预测。表3为本文所用模型的推理时长,本文所提出的ED-DGAT模型推理速度是CS-LSTM模型推理速度的2.61倍。
本文提出了一种基于DGAT建模车辆交互的ED-DGAT模型,用于预测高速公路场景中车辆的未来轨迹。DGAT基于动态注意力机制,有效解决了GAT无法有效捕捉车辆间空间交互的问题。定量试验验证所提出模型的性能,并探究了在模型解码阶段同时建模车辆间交互对预测结果的影响。定性试验结果表明,本文所提出的模型能够预测得到更加真实的车辆未来运动轨迹。NGSIM数据集上的试验结果表明,本文提出的模型与现有轨迹预测方法CS-LSTM相比RMSE评价指标下降了25%,并且推理速度为CS-LSTM模型的2.61倍。
  • *国家重点研发计划项目(2021YFB1600501)
  • 国家自然科学基金项目(61906038)
  • 东南大学“至善青年学者”支持计划项目
  • 中央高校基本科研业务费专项资金项目(2242021R41184)
参考文献 引证文献
排序方式:
[1]
AMMOUN S, NASHASHIBI F. Real Time Trajectory Prediction for Collision Risk Estimation between Vehicles[C]// 2009 IEEE 5th International Conference on Intelligent Computer Communication and Processing. Cluj-Napoca, Romania: IEEE, 2009: 417-422.
[2]
LAUGIER C, PAROMTCHIK I E, PERROLLAZ M, et al. Probabilistic Analysis of Dynamic Scenes and Collision Risks Assessment to Improve Driving Safety[J]. IEEE Intelligent Transportation Systems Magazine, 2011, 3(4): 4-19.
[3]
EIDEHALL A, PETERSSON L. Statistical Threat Assessment for General Road Scenes Using Monte Carlo Sampling[J]. IEEE Transactions on Intelligent Transportation Systems, 2008, 9(1): 137-147.
[4]
AOUDE G S, DESARAJU V R, STEPHENS L H, et al. Driver Behavior Classification at Intersections and Validation on Large Naturalistic Data Set[J]. IEEE Transactions on Intelligent Transportation Systems, 2012, 13(2): 724-736.
[5]
DEO N, TRIVEDI M M. Multi-Modal Trajectory Prediction of Surrounding Vehicles with Maneuver Based LSTMs[C]// 2018 IEEE Intelligent Vehicles Symposium (IV). Changshu, China: IEEE, 2018: 1179-1184.
[6]
ALTCHÉ F, DE LA FORTELLE A. An LSTM Network for Highway Trajectory Prediction[C]// 2017 IEEE 20th International Conference on Intelligent Transportation Systems (ITSC). Yokohama, Japan: IEEE, 2017: 353-359.
[7]
DEO N, TRIVEDI M M. Convolutional Social Pooling for Vehicle Trajectory Prediction[C]// 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW). Salt Lake City, UT, USA: IEEE, 2018: 1468-1476.
[8]
CHEN L, ZHOU Q Y, CAI Y F, et al. CAE-GAN: A Hybrid Model for Vehicle Trajectory Prediction[J]. IET Intelligent Transport Systems, 2022, 16(12): 1682-1696.
[9]
黄玲, 崔躜, 游峰, 等. 适用于多车交互场景的车辆轨迹预测模型[J/OL]. 吉林大学学报(工学版): 1-10( 2022-12-07)[2023-08-21]. https://doi.org/10.13229/j.cnki.jdxbgxb20220728.
HUANG L, CUI Z, YOU F, et al. Vehicle Trajectory Prediction Model for Multi-Vehicle Interaction Scenario[J/OL]. Journal of Jilin University (Engineering and Technology Edition): 1-10( 2022-12-07)[2023-08-21]. https://doi.org/10.13229/j.cnki.jdxbgxb20220728.
[10]
LIANG M, YANG B, HU R, et al. Learning Lane Graph Representations for Motion Forecasting[C]// European Conference on Computer Vision. Glasgow, Scotland, UK: Springer, Cham, 2020: 541-556.
[11]
LI X, YING X W, CHUAH M C. Grip++: Enhanced Graph-Based Interaction-Aware Trajectory Prediction for Autonomous Driving[EB/OL]. ( 2020-05-19)[2023-08-22]. https://arxiv.org/abs/1907.07792.
[12]
GILLES T, SABATINI S, TSISHKOU D, et al. GOHOME:Graph-Oriented Heatmap Output for Future Motion Estimation[C]// 2022 International Conference on Robotics and Automation (ICRA). Philadelphia, PA, USA: IEEE, 2022: 9107-9114.
[13]
赵树恩, 苏天彬, 赵东宇. 基于图神经网络的交互车辆驾驶意图识别及轨迹预测[J]. 汽车技术, 2023(7): 24-30.
ZHAO S E, SU T B, ZHAO D Y. Interactive Vehicle Driving Intention Recognition and Trajectory Prediction Based on Graph Neural Network[J]. Automobile Technology, 2023(7): 24-30
[14]
ZHOU Z K, YE L Y, WANG J P, et al. HiVT: Hierarchical Vector Transformer for Multi-Agent Motion Prediction[C]// 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition. New Orleans, LA, USA: IEEE, 2022: 8823-8833.
[15]
李文礼, 韩迪, 石晓辉, 等. 基于时-空注意力机制的车辆轨迹预测[J]. 中国公路学报, 2023, 36(1): 226-239.
LI W L, HAN D, SHI X H, et al. Vehicle Trajectory Prediction Based on Spatial-Temporal Attention Mechanism[J]. China Journal of Highway and Transport, 2023, 36(1): 226-239.
[16]
连静, 丁荣琪, 李琳辉, 等. 基于图模型和注意力机制的车辆轨迹预测方法[J]. 兵工学报, 2023, 44(7): 2162-2170.
LIAN J, DING R Q, LI L H, et al. Vehicle Trajectory Prediction Method Based on Graph Models and Attention Mechanism[J]. Acta Armamentarii, 2023, 44(7): 2162-2170.
[17]
KOSARAJU V, SADEGHIAN A, MARTÍN-MARTÍN R, et al. Social-BiGAT: Multimodal Trajectory Forecasting Using Bicycle-GAN and Graph Attention Networks[C]// Advances in Neural Information Processing Systems 2019. Vancouver, British Columbia, Canada: IEEE, 2019.
[18]
曾伟良, 陈漪皓, 姚若愚, 等. 时空图注意力网络在交叉口车辆轨迹预测的应用[J]. 计算机科学, 2021, 48(增刊1): 334-341.
ZENG W L, CHEN Y H, YAO R Y, et al. Application of Spatial-Temporal Graph Attention Networks in Trajectory Prediction for Vehicles at Intersections[J]. Computer Science, 2021, 48(S1): 334-341.
[19]
BRODY S, ALON U, YAHAV E. How Attentive Are Graph Attention Networks?[EB/OL]. ( 2022-01-31)[2023-08-22]. https://arxiv.org/abs/2105.14491
[20]
WANG M J, ZHENG D, YE Z H, et al. Deep Graph Library: A Graph-Centric, Highly-Performant Package for Graph Neural Networks[EB/OL]. ( 2020-08-25)[2023-08-22]. https://arxiv.org/abs/1909.01315v2.
[21]
COLYAR J, HALKIAS J. US Highway 101 Dataset: FHWA-HRT-07-030[R]. Washington, DC: U.S. Department of Transportation (USDOT), 2007.
[22]
COLYAR J, HALKIAS J. US Highway I-80 Dataset: FHWA-HRT-07-030[R]. Washington, DC: U.S. Department of Transportation (USDOT), 2007.
[23]
ALAHI A, GOEL K, RAMANATHAN V, et al. Social LSTM: Human Trajectory Prediction in Crowded Spaces[C]// 2016 IEEE Conference on Computer Vision and Pattern Recognition. Las Vegas, NV, USA: IEEE, 2016: 961-971.
[24]
石振周. 基于车道线信息的多目标轨迹预测[D]. 大连: 大连理工大学, 2022.
SHI Z Z. Multi-Objective Trajectory Prediction Based on Lane Information[D]. Dalian: Dalian University of Technology, 2022.
2024年第卷第3期
PDF下载
207
85
引用本文
BibTeX
文章信息
doi: 10.19620/j.cnki.1000-3703.20230582
  • 首发时间:2025-12-23
  • 出版时间:2024-03-24
补充材料
相关文章
文章信息
作者
出版历史
基金
*国家重点研发计划项目(2021YFB1600501)
国家自然科学基金项目(61906038)
东南大学“至善青年学者”支持计划项目
中央高校基本科研业务费专项资金项目(2242021R41184)
作者信息
    东南大学,南京 210096
参考文献
分享链接
https://castjournals.cast.org.cn/joweb/qcjs/CN/10.19620/j.cnki.1000-3703.20230582
分享至
全文二维码

扫描看全文

引用本文
BibTeX
本文的引用情况
2种不同金属材料的力学参数

Family
属数
Number of
genus
种数
Number of
species
占总种数比例
Percentage of
total species (%)

Genus
种数
Number of
species
占总种数比例
Percentage of total
species (%)
鹅膏菌科Amanitaceae 2 11 5.26 鹅膏菌属 Amanita 10 4.78
小菇科 Mycenaceae 2 12 5.74 丝盖伞属 Inocybe 5 2.39
多孔菌科 Polyporaceae 8 14 6.70 蜡蘑属 Laccaria 5 2.39
红菇科 Russulaceae 3 23 11.00 小皮伞属 Marasmius 6 2.87
小菇属 Mycena 11 5.26
光柄菇属 Pluteus 5 2.39
红菇属 Russula 17 8.13
栓菌属 Trametes 5 2.39
关闭全屏