Curb parking occupancy prediction based on real-time fusion of multi-view spatial-temporal information using graph attention gated networks

doi:10.1016/j.asoc.2025.112781

	Curb parking occupancy prediction based on real-time fusion of multi-view spatial-temporal information using graph attention gated networks
	Qian, Chonghui1,2 ; Yang, Kexu1 ; He, Jiangping3 ; Peng, Xiaojing4 ; Huang, Hengjun1,2
	2025-03
在线发表日期	2025-01
发表期刊	APPLIED SOFT COMPUTING
卷号	171
摘要	Effective curb parking management is crucial for reducing traffic congestion, minimizing cruising time, and lowering pollution in smart cities through accurate, timely occupancy predictions. However, traditional spatialtemporal fusion methods often rely on sequential concatenation, leading to spatial information lag and limited real-time fusion. These methods also overlook critical spatial features for curb parking, such as accessibility metrics. To address these limitations, this study introduces the Multi-View Graph Attention Gated Recurrent Unit (MGA-GRU) model, which innovatively fuses spatial-temporal data via graph attention gated networks to enhance curb parking occupancy predictions. Validated on a curb parking dataset from Lanzhou city, China (July 2019 to December 2020), the MGA-GRU model significantly outperforms baseline methods in both short-term and long-term predictions. Ablation experiments demonstrate that each component-the graph attention gates, multi-view graph structure, and real-time fusion-significantly enhances the model's predictive accuracy, highlighting their collective importance in advancing occupancy prediction. By refining spatial-temporal fusion unit, the MGA-GRU model offers vital insights for optimizing curb parking resources and supports efficient traffic management and sustainable urban development.
关键词	Curb parking occupancy prediction Graph attention Gated recurrent unit Data fusion Multi-view graph
DOI	10.1016/j.asoc.2025.112781
收录类别	SCIE ; EI
ISSN	1568-4946
语种	英语
WOS研究方向	Computer Science
WOS类目	Computer Science, Artificial Intelligence ; Computer Science, Interdisciplinary Applications
WOS记录号	WOS:001414667500001
出版者	ELSEVIER
EI入藏号	20250417743502
EI主题词	Traffic congestion
EI分类号	1101 ; 1106.2 ; 432.3 Cargo Highway Transportation ; 662 Automobiles and Smaller Vehicles
原始文献类型	Article
EISSN	1872-9681
引用统计	被引频次[WOS]：0 [WOS记录] [WOS相关记录]
文献类型	期刊论文
条目标识符	http://ir.lzufe.edu.cn/handle/39EH0E1M/38746
专题	统计与数据科学学院工商管理学院信息工程与人工智能学院
通讯作者	Huang, Hengjun
作者单位	1.Lanzhou Univ Finance & Econ, Sch Stat & Data Sci, Lanzhou 730020, Peoples R China; 2.Key Lab Digital Econ & Social Comp Sci Gansu, Lanzhou 730020, Peoples R China; 3.Lanzhou Univ Finance & Econ, Sch Informat Engn & Artificial Intelligence, Lanzhou 730020, Peoples R China; 4.Lanzhou Univ Finance & Econ, Sch Business Adm, Lanzhou 730020, Peoples R China
第一作者单位	统计与数据科学学院
通讯作者单位	统计与数据科学学院
推荐引用方式 GB/T 7714	Qian, Chonghui,Yang, Kexu,He, Jiangping,et al. Curb parking occupancy prediction based on real-time fusion of multi-view spatial-temporal information using graph attention gated networks[J]. APPLIED SOFT COMPUTING,2025,171.
APA	Qian, Chonghui,Yang, Kexu,He, Jiangping,Peng, Xiaojing,&Huang, Hengjun.(2025).Curb parking occupancy prediction based on real-time fusion of multi-view spatial-temporal information using graph attention gated networks.APPLIED SOFT COMPUTING,171.
MLA	Qian, Chonghui,et al."Curb parking occupancy prediction based on real-time fusion of multi-view spatial-temporal information using graph attention gated networks".APPLIED SOFT COMPUTING 171(2025).