2024 Volume 15 Pages 2693-2712
The Passenger Car Unit (PCU) is commonly utilized to convert heterogeneous traffic volume into a homogeneous equivalent, ensuring precise measurement of homogeneous traffic flow and desired service levels in mixed traffic situations in developing countries in general and specifically in Vietnam. Therefore, this present study constructs a new multiple regression equation to estimate the Passenger Car Unit (PCU) for urban roads based on heterogeneous traffic flow components, considering equivalent space mean speed and travel time conditions. From this, evaluate the correlation between two PCU estimation methods: speed-based and multiple linear regression. The new findings show that both methods yield identical PCU estimates at a 2.1-meter car distance. Simultaneously, the study also improved the YOLOv5 model for detecting and counting two-wheelers with high precision, recall, and f1 score. The combination of the regression equation and enhanced YOLOv5 model effectively estimates PCU, expediting the analysis of traffic data for prominent vehicles. A field data survey was conducted on three-lane urban roads in Hanoi, Vietnam. The study results provide specific PCU for urban roads in Vietnam, particularly for two-wheelers, and assess the interrelationship between estimation methods. Recommendations for PCU estimation in heterogeneous traffic flows and potential object detection applications are also discussed to enhance future estimation capabilities.