A Passenger Car and Truck Identification Model Based on the ETC Data of the Expressway in Fujian Province

This Research uses the ETC (Electronic Toll Collection) gantry network toll data of the expressway in Fujian Province to dig into the driving characteristics of each vehicle to establish a vehicle identification model. From time series distribution patterns and vehicle type, significant differences were found between passenger cars and trucks in the number of road changes, the sum of a mile traveled, the sum of journey times, average speed, estimated total mileage, etc. Finally, we adopted logistic regression, decision tree, and random forest models to construct a passenger car and truck identification model. Among them, the random forest model achieves 89.5% accuracy in vehicle type identification based on variables such as the number of line changes, estimated total mileage, average speed, and station entry time. It strongly supports expressway operation management and understanding passenger car and truck travel characteristics.