融合多源时空特征的道路速度预测方法、介质及电子设备
By fusing multi-source features through a multi-input branch network model, the problems of insufficient multi-source feature fusion and inadequate road network association modeling in existing technologies are solved, achieving high-precision prediction of road speed in port areas, especially maintaining stable prediction results in highly dynamic and complex scenarios.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- SHANGHAI INTERNATIONAL PORT
- Filing Date
- 2026-06-18
- Publication Date
- 2026-07-17
AI Technical Summary
Existing road speed prediction methods suffer from insufficient multi-source feature fusion, inadequate road network correlation modeling, and low prediction accuracy, especially in highly dynamic and complex scenarios such as port areas.
A multi-input branch network model is adopted, which integrates road speed, time, weather and road network correlation features. It processes multi-dimensional features through spatial attention weights and long short-term memory networks to achieve deep fusion and feature splicing of multi-source data and output multi-step prediction values.
It improves the accuracy and stability of road speed prediction in port areas, maintains stable prediction accuracy in highly dynamic and complex scenarios, and has good scene transfer capabilities.
Smart Images
Figure CN122416751A_ABST