基于门架数据的高速公路货车流量短时预测

打开文本图片集

中图分类号：TP183 文献标志码：A DOI：10.13705/j. issn.1671-6841. 2024025

文章编号：1671-6841（2025）06-0058-07

Abstract： Highway freight always occupy a large share in the freight system. Compared with other traffc sources，data colected from gantry were more accurate. But the data were difficult to obtain，so the existing forecasting models rarely used gantry data to predict highway freight traffic.To address this issue，a short-term prediction model for highway freight traffic based on gantry data was proposed. Initially，the highway freight data were preprocessed. Then，an integration of atention mechanisms with AGCN was employed to excavate spatial correlations within the data，while ResNet and LSTM were utilized to uncover temporal dependencies.Finally，feature fusion was applied to derive the predicted highway freight traffic results.By comparative experiments，it was demonstrated that the proposed model exhibited higher accuracy in short-term highway freight traffic forecasting compared to baseline models such as LSTM and STNN.

Key words： short-time flow prediction; gantry data;deep learning; residual neural network ; long shortterm memory network

0 引言

随着我国经济的不断发展，货物运输需求不断增加，其中公路货运在货运体系中持续占据重要地位。（剩余10010字）