Using Simulated Data, A Cloud-Based Traffic Management System with V2X-Enabled Dynamic Traffic Control Outperformed Existing Models, Achieving a 25 Percent Improvement in Traffic-Flow Efficiency and a 40 Percent Faster Response Time.
Model Framework Enhanced Data Responsiveness to Improve Simulated Travel Times and Optimize Traffic Flow.
Canada
Integrating V2X solutions in intelligent green cities: an AI-driven point exchange system approach
Summary Information
Vehicle-to-Everything (V2X) technologies use wireless communications to link vehicles, mobile devices, and roadside infrastructure, enabling decisions that improve roadway safety, mobility, and operational efficiency. This study introduced a cloud-based traffic management system (TMS) that leverages a four-tier architecture (IoT, fog, cloud, application) for real-time traffic optimization. The fog layer served as an intermediate data processing and decision-making stage, which, given the substantial volume of data generated by V2X-enabled vehicles, was critical in managing the data efficiently.
METHODOLOGY
The study used seven years’ worth of vehicle data from the Canadian government’s open data portal to build a prediction model. It also developed a traffic signal control algorithm that uses V2X data and AI-driven insights to dynamically adjust signal timing. The model framework was organized into four layers:
- Application layer: Interfaces and applications directly accessible to users.
- Internet of Things (IoT) layer: Sensors and devices collect data in real time.
- Fog layer: Intermediate processing points enhancing data responsiveness.
- Cloud layer: Centralized data storage and advanced processing capabilities.
The evaluation relied on simulated traffic and V2X data and was assessed by running scenarios and comparing results against baseline methods (e.g., traditional AI models such as support vector regression (SVR)), focusing on metrics such as latency, response time, and traffic efficiency.
FINDINGS
A comparative analysis of the best-performing traditional model (SVR) to the cloud-based TMS model showed better performance of the latter across all evaluated metrics.
- Traffic Flow Efficiency: a 25 percent increase in traffic flow efficiency, compared to 10 percent improvement by SVR.
- Latency Reduction: a 30 percent reduction in latency, outperforming SVR’s 15 percent.
- Response Time Improvement: a 40 percent improvement in response time, compared to 20 percent by the SVR.
