Research on traffic state prediction method based on traffic flow prediction under multi-time granularity
Abstract Accurate traffic state prediction is of great significance to effective traffic planning and management. The different characteristics of different sections and time-varying traffic patterns pose a challenge to the accurate prediction of traffic conditions. Most of the existing traffic stat...
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Nature Portfolio
2025-07-01
|
| Series: | Scientific Reports |
| Subjects: | |
| Online Access: | https://doi.org/10.1038/s41598-025-10267-9 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Abstract Accurate traffic state prediction is of great significance to effective traffic planning and management. The different characteristics of different sections and time-varying traffic patterns pose a challenge to the accurate prediction of traffic conditions. Most of the existing traffic state prediction researches directly use historical traffic flow data or current traffic state in a single time granularity to predict future traffic state, which is difficult to capture the nonlinear characteristics of complex traffic flow effectively. In order to improve the accuracy of traffic state prediction model, a traffic state prediction method based on multi-time granularity traffic flow prediction is proposed. Firstly, in order to improve the interpretability and predictability of data, a Seq2Seq traffic flow prediction model based on multi-parameter fusion is proposed. Secondly, the traffic state identification method based on temporal and spatial attributes is studied. Then, based on the prediction results of traffic flow parameters, an indirect prediction method of traffic state based on multi-time granularity is designed. Finally, based on the experimental results, it is verified that the traffic state prediction method based on multi-time granularity has higher prediction accuracy, and the effectiveness of the indirect prediction method based on the traffic flow prediction results is verified based on different “direct prediction” models. |
|---|---|
| ISSN: | 2045-2322 |