Dilemma zones on the approach to high-speed signalized intersections have been identified as a safety problem that can contribute to rear–end and right–angle crashes. However, in situations when a pair of vehicles is caught in the dilemma zone, it is not well understood how the following vehicle’s headway and classification influence the leading driver’s decision in response to the Circular Yellow (CY) indication. This complex vehicle-vehicle-intersection interaction, and the consequences of faulty driver decision making warrant further evaluation.
This study analyzed driver behavior at the onset of the CY indication at isolated high-speed signalized intersections and evaluated three elements of the challenge: 1) yellow light laws and driver training manual instructions, 2) driver’s visual attention, and 3) driver’s stop-go decisions.
Findings indicated inconsistencies between the Driver Training Manual (DTM) guidance and yellow state laws. 74% of states followed a Class 1 (permissive law) yellow law while 72% of states followed Class 3 (restrictive law) guidance in DTMs. This inconsistency between state yellow laws and DTM guidance may contribute to inconsistencies in driver comprehension and decision making in response to CY indications, particularly when drivers travers state boundaries.
The highest percentage of Total Fixation Duration (TFD) was allocated to the traffic signal head (78.4%), followed by the rear view (20.3%) and then side view mirrors (1.3%). Repeated-measures ANOVA was performed to determine whether TFD differed between scenarios for each Area of Interest (AOI). When a significant effect was observed, pairwise comparisons were conducted to find the origin of the difference. For the traffic signal AOI, the following vehicle type and Time to Stop Line (TTSL) had significant effects on the TFD. For the rear-view mirror AOI, the following vehicle type and time headway had significant effects on TFD when the driver looked at the rear view mirror. No significant effect was observed for any independent variable or the two or three-way interactions on the TFD on the side view mirrors. Random-effect Tobit regression was used as the data involved repeated measures and numerous zeros. Two random-effect Tobit regression models were developed to deal with two AOIs (traffic signal and rear view mirror). TTSL, driver age of 45–55 years, and vehicle speed at the onset of CY indication (35–45) mph had positive and statistically significant effects on the TFD for the traffic signal with a 95% confidence interval (CI). For the rear view mirror, TTSL, driver age of 55–65 years, and drivers with some high school or less all had positively significant effects for TFD while time headway, driving once per week, and driving a van in one’s personal life had negatively significant effects for the TFD for a 95% CI.
Finally, decision making results indicated that 51% of drivers decided to stop prior the stop line while 49 % chose to proceed through an intersection in response to CY indication. Nearly all drivers (97%) went through an intersection when they were 2.5 seconds from the stop line and red-light running violations start to increase when TTSL was 5.5 seconds. A random parameter binary logit model was used to deal with unobserved heterogeneity across observations. A total of 4 parameters were found to be statistically significant on driver’s decision making at the onset of CY indication. The findings indicated that TTSL, time headway, and driver age of 20–36 years increased the probability of stopping while vehicle speed at the onset of the CY indication decreased the probability of stopping. The results showed that following vehicle type did not influence driver’s decision to proceed or stop in response to the CY indication with statistical significance.
Funding Statement (additional comments about funding)
I would like to thank the Iraqi Ministry of Higher Education and Scientific Research as well as the University of Anbar for providing me the opportunity and support to pursue my Ph.D. degree from Oregon State University. Without their support, this rewarding educational experience would not have been possible.