A Comparitive Analysis of Stop/Go Driving Behavior during the Amber Light at City Signalized Intersecdtions

Abstract

This paper investigates the relationship between the driver’s stop/go behavior when encountering signal change from green to amber and traffic parameters of each car arriving at the intersection during the amber light. The data was collected from two intersections (one is high speed the other is low speed) both located in Beijing with video camera for a period of two weeks. 462 samples are extracted from the video using the software Adobe Premiere Pro CS4. Each sample include the stop/go decision, vehicle’s yellow onset speeds, vehicle’s yellow onset distance from the intersection, position (leading/following) in the traffic flow, vehicle’s speed, amber light time when vehicle travel through the stopline. The line regression technology and decision tree model are respectively used to analyse the difference of influence of traffic parameters to stop/go driving behavior between the high speed and low speed intersections. And it is clearly observed that the dilemma zone at low speed intersection is a little wider than that at high speed intersection althouth the averge arriving speed is lower. And the decision tree model result shows that parameter location play the most important role in sotp/go decisions, and parameter speed only works when the car falls into dilemma zone.