استجابة السائق للتوقيت الأصفر عند التقاطعات ذات الإشارة المختارة في مدينة كربلاء, العراق

Driver Response to Yellow Timing at Selected Signalized Intersections in Karbala City, Iraq

Abstract

The city of Karbala in Iraq is one of the most important cities that is witnessing a significant increase in its population. Moreover, due to the importance of the city from a religious perspective, it has a large number of visitors during the year. Therefore, the study of driving behavior at signalized intersections in this city can provide an analytical background for a safer traffic signals design that can reduce the probability of vehicle accidents at intersections. In this context, this work is an experimental study that is aimed at analyzing driving behavior and estimating the driver’s decision to move or stop at signalized intersections when the yellow light is presented. In particular, in this period, the driver may not be able to safely stop before the stop line, nor crosses the intersection. This is known as the dilemma zone, which is considered critically important to understand and characterize, yet the estimation of the driver’s decision at this zone is considered highly stochastic because the driver’s decision can be changed from one driver to another. This is a data-driven approach that ends up providing a model that can be used to estimate the driver’s decision from observed measurements. In this experimental study, the input variable xi consists of four features; speed, headway, TTSL (time to stop line), and vehicle type. The output variable yi is binary (takes either 0 or 1) that represents the driver’s decision to pass (yi = 0) or not pass (yi = 1) the intersection. The data were collected by installing cameras placed at the Al- Dhareeba intersection and Saif Saad intersection for five workdays in Karbala city. The passing and stopping vehicles were calculated for the yellow light, as well as those vehicles violating the traffic signal during traffic intersections for the beginning of the red light. In this context, Red Light Running (RLR) is one of the most dangerous riding behaviors at mixed traffic intersections and one of the most serious safety issues. In the first set of experiments, RLR violations were measured for five days from 8:00 a.m.-10:00 p.m. According to these measurements, there was a high percentage of RLR violations that occurred in both signalized intersections. During the hours of 8:00-9:00 a.m., the Al- Dhareeba intersection had the highest percentage of RLR violations (67.60%), while Saif Saad intersection had 45.46%. In the second set of experiments, the input and output variables {(xi , yi)} were measured for five days during rush hours 8:00 am-10:00 a.m. and 5:00 p.m.-7:00 p.m. This set of experiments provided 513 samples (xi , yi) for each intersection (a total of 1026 samples for both intersections). A logistic regression binary model is used in this study. To train and test this classifier, SPSS Statistics has been used in this study. The results for Al-Dhareeba intersection indicated that there was an insignificant relationship between the speed and headway and vehicle type with the driver’s decision to go or stop at the yellow light, while the TTSL feature is significant with the driver’s decision having a p-value of 0.01. The negative sign in the coefficient (β) indicated that the probability of stopping increases with the decrease of the speed and headway. In Saif Saad intersection case, the results indicated a significant relationship between the speed and TTSL with the driver’s decision. The negative sign in the coefficient (β) indicated that the probability of stopping increases with the decrease in headway. In the third set of experiments, data were collected using electronic questionnaire questions open to all drivers on the internet, where a statistical analysis has been carried out to provide a qualitative assessment of the effect of gender, age, and education level on the driver’s decision.