Application of Ensemble Learning and VISSIM in Intersection Traffic Flow Prediction and Signal Timing Optimization
The intersection, as the focal point of urban traffic networks, is critical not only for connecting roads, but also for traffic control and optimization. This research integrates ensemble and reinforcement learning with the VISSIM simulation tool to investigate traffic flow prediction and signal tim...
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2024-01-01
|
| Series: | IEEE Access |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/10802895/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | The intersection, as the focal point of urban traffic networks, is critical not only for connecting roads, but also for traffic control and optimization. This research integrates ensemble and reinforcement learning with the VISSIM simulation tool to investigate traffic flow prediction and signal timing at intersections. Initially, an ensemble learning algorithm accurately predicts intersection traffic flow. The SARSA-A2C reinforcement learning technique is then used in conjunction with VISSIM to optimize signal timing, reducing delays and congestion. By integrating the predictions with the SARSA-A2C algorithm, a hybrid strategy for predictive signal timing optimization is implemented. Using four consecutive intersections on Shanghai’s West Yan’an Road as a case study, the results, validated by model and network performance metrics, demonstrate the efficacy of the hybrid strategy: a 20.0% decrease in average stops per vehicle, an 11.7% reduction in total delay, a 15.8% decrease in total stop delay, and a 25.0% reduction in average vehicle delay. This study presents an innovative approach to urban traffic management, with significant theoretical value and promising application potential. |
|---|---|
| ISSN: | 2169-3536 |