Modeling Drivers' Behavior in Selecting Stop/Go at the Onset of Red-Phase at Signalized Intersections (Case Study of Qazvin City)

Red-light running (RLR) is considered as a highly dangerous driving act. Reducing this risky behavior depends on understanding its prevalence, as well as the drivers which are involved. This study investigates the video images of traffic cameras and central smart program in four intersections of Qazvin. Peak and off-peak conditions were assessed in either sunny or rainy weather and drivers' behavior in red times was studied using binary logit model. It has been considered that red light running is often occurred unintentionally. Generally, according to the observations, 96 percent of red phase violations are happened during the first two seconds. Recent studies have concluded that drivers’ decision to stop or cross the red light is complex but can be reasonably predicted based on numerous factors (e.g. number of pedestrians in streets and time passed in red phase, etc). Vehicle’s speed is another significant factor which highly effects the crossing probability during red phase. Moreover, in situations where vehicles` distance from the intersection is more than 20 m at the onset of a yellow-phase, the passing probabilities are zero percent and if vehicles’ distance from the intersection is more than 9 m, red light running probability will be 51 percent (Other variables considered as their average).