农业装备与车辆工程2025，Vol.63Issue(9)：123-128,6.DOI:10.3969/j.issn.1673-3142.2025.09.022
融合车型特征的高速公路短时交通量预测研究
Research on short-term traffic volume prediction of expressways integrating vehicle type characteristics
刘宇可 1李慧2
作者信息
- 1. 西华大学汽车与交通学院,四川 成都 610039
- 2. 西华大学汽车与交通学院,四川 成都 610039||四川西华交通司法鉴定中心,四川 成都 610039
- 折叠
摘要
Abstract
To improve the prediction accuracy of short-term traffic flow on expressways,a GCN-Transformer prediction model integrating spatiotemporal characteristics was constructed.In view of the characteristics of expressway traffic flow,such as diverse vehicle types,high driving speeds,and significant speed differences between different vehicle types,and to further reveal the impact of vehicle type characteristics on traffic flow,the proportion of vehicle types was introduced into the prediction model as a key feature.Empirical results showed that the heterogeneity of vehicle types had a significant impact on short-term traffic flow prediction,and the proposed GCN-Transformer model could effectively improve prediction accuracy.After introducing vehicle type features,the model performance was further improved:the Mean Absolute Error(MAE)and Root Mean Square Error(RMSE)were reduced by 0.12 and 0.47 respectively,and the prediction fitting degree was increased by 1.64%.关键词
短时交通量预测/车型特征/高速公路/深度学习/GCN—Transformer模型Key words
short-term traffic volume prediction/vehicle model features/expressway/deep learning/GCN-Transformer model分类
交通工程引用本文复制引用
刘宇可,李慧..融合车型特征的高速公路短时交通量预测研究[J].农业装备与车辆工程,2025,63(9):123-128,6.
