一种基于稀疏观测数据的交通枢纽客流时空分布预测方法
By constructing the PD-STHGAT deep learning framework and combining hypergraph attention network and gated linear unit, the problem of passenger flow prediction in transportation hubs under sparse observation data is solved, and efficient and accurate prediction of passenger flow distribution in large-scale integrated transportation hubs is achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- SOUTHEAST UNIV
- Filing Date
- 2025-10-13
- Publication Date
- 2026-07-17
AI Technical Summary
Existing technologies are difficult to effectively predict passenger flow distribution in large-scale integrated transportation hubs, especially under conditions of sparse observation data, and are difficult to generalize and cannot capture high-order spatiotemporal correlations.
We construct the PD-STHGAT deep learning framework based on pedestrian dynamics knowledge and high-order hypergraph modeling methods. Combining hypergraph attention network and gated linear unit, we achieve passenger flow prediction by interpolating sparse passenger flow distribution matrix and introduce reconstruction loss and physical loss to constrain the prediction results.
It enables accurate prediction of passenger flow distribution in transportation hubs under sparse observation data conditions, captures high-order spatiotemporal characteristics, and provides a scientific basis for management and control.
Smart Images
Figure CN121305865B_ABST