Traffic flow prediction model based on multi head dynamic attention and multi gate time convolution

GuangBing Bao; Jinyuan Yang; XiaoLian Wu; Zhonghao Liu; Jianhang Zhang

doi:10.1117/12.2668601

16 February 2023 Traffic flow prediction model based on multi head dynamic attention and multi gate time convolution

GuangBing Bao, Jinyuan Yang, XiaoLian Wu, Zhonghao Liu, Jianhang Zhang

Proceedings Volume 12591, Sixth International Conference on Traffic Engineering and Transportation System (ICTETS 2022); 125913P (2023) https://doi.org/10.1117/12.2668601
Event: 6th International Conference on Traffic Engineering and Transportation System (ICTETS 2022), 2022, Guangzhou, China

Abstract

In the era of rapid development of intelligent transport construction, predicting traffic flow is a crucial aspect of managing and operating contemporary transportation networks. However, traffic prediction tasks have complex spatio-temporal connections. On the one hand, the temporal characteristics of different time scales affect the accurate prediction of traffic flows, and on the other hand, the potential spatial relationships between road traffic cannot be reasonably expressed by the current predefined graph structure. In order to anticipate traffic flow, this article suggests a model (Dynamic attention and Multi gated time convolution Network) based on dynamic attention and multiple gated temporal convolution. The model introduces a multi-gated time convolution layer to extract temporal features of different time ranges and extracts potential spatial relationships between traffic nodes through a multi-headed dynamic attention mechanism to dynamically acquire spatio-temporal variations. Unlike existing models, the model does not require a priori road knowledge of the traffic map, but instead adjusts the inter-node relationships through multi-head dynamic attention. At the same time, the convolution kernel's size is expanded by the model and the gating output of the temporal convolution on top of the original temporal gating to improve the accurate prediction of different temporal pattern tasks. The model forecasts greater properly than the contrasted baseline methodologies, according to analytical outcomes.

Citation Download Citation

GuangBing Bao, Jinyuan Yang, XiaoLian Wu, Zhonghao Liu, and Jianhang Zhang "Traffic flow prediction model based on multi head dynamic attention and multi gate time convolution", Proc. SPIE 12591, Sixth International Conference on Traffic Engineering and Transportation System (ICTETS 2022), 125913P (16 February 2023); https://doi.org/10.1117/12.2668601

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available

Members: $17.00

Non-members: $21.00 ADD TO CART

PROCEEDINGS
8 PAGES

DOWNLOAD PAPER SAVE TO MY LIBRARY

GET CITATION

RIGHTS & PERMISSIONS

Get copyright permission Get copyright permission on Copyright Marketplace

KEYWORDS

Convolution

Data modeling

Networks

Roads

Feature extraction

Performance modeling

Head

Show All Keywords

Keywords/Phrases

Search In:

Publication Years