Estimating time-varying queue length at signalized intersection by using color-coded traffic on Google Maps

Abstract

Traffic jam, at a signalized intersection, is a common problem in urban areas that emergency vehicles such as ambulances are usually stuck in a queue. Thus, emergency medical transportation may be impeded leading to unnecessary damage or death. This research aims to develop an intelligent traffic signal system that prioritizes emergency vehicles. The main focus is developing a method for measuring the time-varying queue length at a signalized intersection, which is an integral part of the system's operation. In the past, most of research analyzed the queue lengths by using traffic detectors, but they have rarely been installed and used in Thailand In this paper, we present a new approach to estimate the queue length that does not need a traffic detector; instead, it utilized color-coded traffic data from Google Maps to predict the queue length based on three different methods: which are Multiple Linear Regression, Random Forest, and The Mean method. The study found that the length of the colored bars displayed on Google Maps can, to some extent, provide a good approximation of the actual queue length trend, and are more correlated with two methods especially in rush hours with heavy traffic.  The Random Forest, among three methods, yields the best performance in predicting the queue length when compared to the other two methods.