Institute of Transportation, MOTC

       Intersections are hot spots for traffic crashes, especially during the amber phases, which is a transitional period for changing the right of way of different traffic flow. Drivers will face the choice of passing or stopping. Once a wrong decision is made, the risk of crash will increase. This study uses radar and video data to observe the vehicles that travel through the intersection during amber periods. Variables considered include the vehicle speed, vehicle position at the onset of amber light, gap between the front and the rear vehicles. Traffic crash data was used to define the three improper driving behaviors at the intersection accident, which were identified as speeding, not maintaining enough distance, and without completely passing the intersection when red is on. Vehicles with at least one improper driving behavior were simulated in accordance with site speed limit and appropriate acceleration (2.56 m/s2) for passing, and deceleration (3 m/s2) for stopping. The vehicles that continue to travel at the intersection during the amber are classified into three categories: correct decision, under dilemma zone, and improper decision. Through logistic regression and discriminant analysis, two models were developed to correlate factors affecting driver’s decision. It was found that about 20% vehicles red light running, and 8% vehicles both red light running and speeding. In addition, it was found that vehicle distance to stop line, initial vehicle speed, length of amber phase, width of crossing road, vehicle length, green arrow with red light are factors significantly affect driving decisions