Automatic Positioning Method for Expressway Gantry by Mining Electronic Toll Collection Data

ABSTRACT To address the issue of inaccurate or missing gantry position information, which severely impacts Electronic Toll Collection (ETC) system efficiency and data utility, this work proposes a novel method for automatic gantry positioning based on ETC transaction data. Firstly, a mileage generation model (MGM) of the expressway gantry section was constructed. Secondly, the gantry section local and road section global outlier data cleaning method (ODC) was designed. Thirdly, the problem of sparseness of road point elements was solved through densification, and then an ETC gantry position automatic generation (GPG) algorithm based on dichotomy optimization was designed. Finally, the method was verified by field experiments with expressway ETC transaction data. Theoretical analysis and experimental results indicate that the data cleaned by the ODC algorithm, which presents a Gaussian distribution, is suitable for the MGM model. Meanwhile, utilizing the data samples in the early morning hours, the average error of the segment mileage generated by the MGM model has reached the accuracy requirement of 10 m. Further, the geographic coordinates of the gantry obtained through the GPG algorithm are all located within the roadway, and the positioning errors of the gantries achieve consistent accuracy and independence. The work shows that the constructed model utilizes ETC data to realize accurate and efficient automatic positioning of the ETC gantry. It can not only be used to verify the position information of the existing gantry but also realize the automatic generation of the position of the gantry with missing position and the new gantry in the future.