Drivers' mistakes in making immediate decision facing yellow signal interval to stop or go through the intersection is one of main factors contributing to intersection's safety. Incorrect decision might lead to a red light running and a right-angle Collison when passing through the intersection or a rear-end collision when failing to stop safely.Improperly timed traffic signal intervals result in the inability of the drivers to make the right decision and can place them in the dilemma zone. Advance warning systems (AWS) have been used to provide information about the downstream traffic signal change prior to approaching the intersection. On the other hand, advance warning systems increase drivers approach speed according to the literature. However, effect of AWS on dilemma zone has not been studied before. The goal of this thesis is to minimize the number of vehicles caught in dilemma zone by determining more precise boundaries for dilemma zone and to reduce the number of red light violations by predicting the red light runners before arriving to the intersection. Here, dilemma zone boundaries at the presence of AWS has been reexamined with the aid of a large dataset (more than 1870 hours of data for two different intersections). Upper dilemma zone boundaries found to be higher for the intersections with AWS. This is due to vehicles' increasing the speed at the flashing yellow sings to escape the dilemma zone.Moreover, an algorithm for predicting red light runners and distinguishing them from right turners is presented.