Estimating of Origin-Destination Matrix on Expressway Using the Automatic License Plate Recognition and Electronic Toll Collection Data

Abstract

Origin-Destination (OD) matrix of traffic volume is the crucial information for network traffic planning and analysis. Traditionally, gathering OD data gathering OD data involved surveying road users, which has limitations in terms of the limited number of respondents and the inability to identify individual vehicles. The accuracy of the resulting origin-destination matrix might not be sufficient. In the recent years, the automatic vehicle identification (AVI), such as optical character recognition (OCR) technology and dedicated short- range communication (DSRC), has advanced significantly. Automatic license plate recognition (ALPR) systems can accurately read license plates and match the origin and destination points of vehicles. In addition, the penetration rate for the electronic toll collection (ETC) in Thailand has continuously grown. Therefore, this study aims to implement ALPR systems and develop an algorithm for estimating the origin-destination matrix of traffic volume using data from ALPR system. This algorithm calculates the proportion of origin-destination patterns and incorporates it with traffic volume data obtained from toll systems. The performance of the origin-destination matrix is assessed using data from the screenline sensors located on the expressway. The findings indicate that the method for determining the origin-destination matrix of traffic volume based on ALPR is accurate and can be utilized in other applications. The use of ETC data to enhance the OD estimation yielded mixed results, i.e., higher accuracy in some routes and lower accuracy in the other routes.