Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
70 C H A P T E R 5 Conclusions and Discussion 5.1 Conclusions The objective of this research was to analyze driver behavior and driver comprehension impacts of retrofitting existing three- and five-section signal displays with the FYA indication. This objective was accomplished by conducting a computer-based static study at various locations in Madison, Wisconsin and Amherst, Massachusetts, and a dynamic driving simulator study at the same locations. The results from the studies were analyzed to determine if the location of the FYA indication within any given signal display arrangement had a significant impact on driver comprehension. In addition, other variables were analyzed to determine if they had any impact on driver comprehension. These variables included the signal display arrangement, the thru indication, the presence of opposing traffic, and demographics. Note that there is not enough sensitivity in the simulator to evaluate issues pertaining to the placement of separate versus shared left-turn signal displays and tightly spaced (less than 8 feet) signal displays. For all signal display arrangements evaluated, there was not a statistically significant difference in driver comprehension when the FYA indication was located in the middle section or the bottom section of the signal display. In Wisconsin, scenarios involving a FYA indication in the middle section had a correct response rate of 77.4%, while scenarios involving a FYA indication in the bottom section had a correct response rate of 78.1%. In Massachusetts, the correct response rates were 68.1% and 70.0%, respectively. Analyses were also completed to determine whether the location of the FYA indication was comprehended differently for selected signal displays and thru movement indications. There were no significant differences identified. When considering the traffic signal display arrangement, driver comprehension was the lowest for the five-section cluster signal display with simultaneously illuminated left-turn and thru movement indications. The fact that the five-section cluster signal display with simultaneous indication had the lowest correct response rate in both of the states is not surprising, as previous research completed by Noyce and Kacir, and Knodler, et al., had shown that driver comprehension is significantly reduced with simultaneous indications (Noyce and Kacir 2005, Knodler, Noyce, Kacir, and Brehmer 2005). Overall, the results show that the signal display arrangement did not impact driver comprehension of the FYA indication. Driver comprehension was significantly impacted by the adjacent thru indication in Wisconsin. Left-turn signal displays with the FYA indication with a CG thru movement indication had the highest correct response rate at 89.7%, followed by the CY indication with a correct response rate of 74.6%. When the adjacent indication was a CR indication, the lowest correct response rate at 66.9% was observed. The differences between each of these adjacent thru indications was statistically significant. Massachusetts had similar results with correct response rates of 79.1% with a thru CG indication, 69.9% with a thru CY indication, and 60.8% with a thru CR indication, respectively. Results show that the thru indication is considered in left-turn decision making and has a large impact on driver comprehension.
71 A total of 56 drivers participated in the driving simulator study, providing 466 scenario responses. A bimodal FYA indication in a three-, and five-section signal display was featured in 339 of the scenarios evaluated. Eye tracking data was also collected for 16 drivers. Findings show no significant differences in driver comprehension when the FYA indication was presented bimodally in the bottom section or bimodally in the middle section of a three-section vertical signal display. Post-driving simulator participant queries found that 38% preferred the FYA indication in the middle section, 9% preferred the bottom, and 54% indicated no preference. There was also no significant difference in correct responses between the three-section and four- section vertical signal displays. Driver comprehension was significantly lower when the FYA indication was added bimodally to the five-section cluster signal display and the simultaneous thru movement indication was also illuminated. This result was especially true when the simultaneous and thru movement indication was CR. Speed trajectory data showed that drivers decreased their speed and/or stopped to wait for an acceptable gap in each of the experimental scenarios. Trajectories also showed vehicles stopping past the stop bar of the approach and then proceeding after selecting a gap. In multiple instances, drivers appeared to make a decision to proceed and continued moving, only to reconsider their decision by slowing to yield to opposing traffic and search for a gap. Observed behavior was expected when drivers were presented with an FYA indication. There was evidence to support the observation of driver confusion associated with the five-section cluster signal display and simultaneous CR thru indication. Average speed was lower and speed variance less at the stop bar location, indicating a greater number of drivers hesitated and/or stopped as they tried to comprehend the appropriate action. No differences in driving behavior were observed in the trajectories of scenarios with a bimodal indication and of those without the bimodal indication. Furthermore, no difference in the trajectories was observed when comparing those scenarios in which a RA indication was presented prior to the signal going into the FYA indication mode. Eye tracking data showed that participants spent more time looking at opposing traffic than at the combination of FYA and thru movement indications. Evidence of a more rapid search process between opposing traffic and the signal displays was observed when drivers did not completely comprehend the traffic signal displays. This evidence was clearly observed with the five-section cluster signal display with a simultaneous CR indication. Combined research results show that the FYA indication can be effectively used (retrofit) in existing three-section vertical signal displays, when combined bimodally within the YA or GA indication section, without negatively impacting driver behavior and comprehension. The FYA indication retrofit is not recommended for use in any section of a five-section cluster signal display with or without simultaneous indications. Other signal display arrangements do not impact driver comprehension of the FYA indication, as long as simultaneous indications are not used. Although the static and driving simulator studies do not show a difference in driver behavior and driver comprehension when the FYA indication is either in the middle or bottom section of the three-section signal display, the post-driving simulator questionnaire clearly showed that drivers prefer the FYA/YA bimodal signal indication within the middle section. Furthermore, with the FYA/GA bimodal indication in the âgoâ section of the signal display, the probability of a fail- critical error was increased. Recall that reduction in fail-critical errors has proven to be the significant contribution of using a FYA permissive left-turn indication. Although there is some concern with the FYA indication being bimodal with the YA indication in drivers detecting the
72 change from FYA to solid YA during the change interval, post-driving simulator participant questionnaire responses, along with the high probability of a fail-safe response and no observed concerns in the experimental evaluations, makes the FYA/YA bimodal indication the most desirable retrofit. Therefore, researchers recommend that the FYA indication can be effectively used in a three-section traffic signal display only when used bimodally with the steady YA indication. Speed trajectory data found that drivers decrease their speed and/or stopped to wait for an acceptable gap in each of the experimental scenarios. Trajectories also show vehicles stopping past the stop bar of the approach and then proceeding after selecting a gap. In multiple instances, drivers appeared to make a decision to proceed and continued moving, only to reconsider their decision by slowing to yield to opposing traffic and search for a gap. Observed behavior was expected when drivers were presented with a FYA indication. There was evidence to support the observation of driver confusion associated with the five-section cluster signal display and simultaneous CR thru indication. Average speed was lower and speed variance less at the stop bar location, indicating a great numbers of drivers who hesitated and/or stopped as they tried to comprehend the appropriate action. No differences in driving behavior were observed in the trajectories of scenarios with a bimodal indication and of those without the bimodal indication. Furthermore, no difference in the trajectories was observed when comparing those scenarios in which a RA indication was presented prior to the signal going into the FYA indication mode. Eye tracking data showed that participants spent more time looking at opposing traffic than at the combination of FYA and thru movement indications. Evidence of a more rapid search process between opposing traffic and the signal displays was observed when drivers did not completely comprehend the traffic signal displays. This evidence was clearly observed with the five-section cluster signal display with a simultaneous CR indication. Research results show that the FYA indication is an acceptable retrofit for existing three-section vertical signal displays when combined bimodally with the YA or GA indication section. The FYA indication retrofit is not recommended for use in any section of a five-section cluster signal display. The signal display arrangement does not impact driver comprehension of the FYA indication, as long as simultaneous thru movement indications are not used. 5.2 Discussion Although, the static study and driving simulator study do not show a difference in driver comprehension and driver behavior when the FYA indication is either the middle or bottom of the three-section signal display, the post-driving simulator questionnaire presents evidence to suggest that drivers prefer the FYA/YA bimodal indication in PPLT signal displays. Additionally, with the FYA/GA bimodal indication is in the âgoâ section of the signal display, the probability of a fail-critical error is increased. Reduction in fail-critical errors has proven to be the significant contribution of using a FYA permissive left-turn indication. Although there is some concern with the FYA indication being bimodal with the YA indication in driving detecting the change from FYA to solid YA during the change interval, post-driving simulator participant query responses, along with the high probability of a fail-safe response, makes the FYA/YA bimodal indication the most desirable retrofit. Therefore, researchers recommend that the FYA indication can be effectively used in a three-section traffic signal display only when used bimodally with the YA indication.
