Queue Length Estimation for Signalized Intersections Using License Plate Recognition Data

Queue length is an important variable in evaluating the performance of traffic management systems, especially under congested conditions. This problem has been studied in the literature based on different kinds of traffic sensors, such as loop detectors and other mobile sensors. Automatic Number Plate Recognition systems and Automatic Vehicle Identification systems have received significant attentions due to their active deployments in recent years, which provides massive License Plate Recognition (LPR) data sources. In this paper, new methods for queue length estimation at signalized intersections are proposed using LPR data and signal schemes, with the key point being the intrinsic connections between travel time of individual vehicles and queue composition in each cycle. The queue length in the previous cycle is calculated by considering the detailed trajectories of individual vehicles. Due to the hysteresis of the queue length in the previous cycle, the queue length in the immediate past cycle is formulated as a prediction problem using regression analysis. The algorithms were evaluated using field data of Guiyang city in China, and the proposed models? results fully coincide with the experimental findings even in congested conditions.