Transportation Research Record: Journal of the Transportation Research Board

The lack of a national standard, recommended practice, or set of guidelines for determining the duration of the yellow change and red clearance intervals for the timing of traffic signals has left each transportation agency to determine its own practices. The objective of NCHRP Project 03–95 was to develop a comprehensive and uniform set of recommended guidelines for determining safe and operationally efficient yellow and red intervals at signalized intersections. To accomplish this objective, the study established a state of knowledge and a state of practice through a review of existing guidelines and various literature sources and a survey of transportation agency practitioners. The research also conducted a comprehensive national investigation of driver behavioral characteristics through an extensive field data collection and analysis effort. These efforts confirmed the accepted values for perception–reaction time (1.0 s) and deceleration rate (10 ft/s²) and established 85th percentile speed estimations for through and left-turning vehicles. Justification was also provided for accounting for start-up delay. From the findings, a succinct recommended guideline for the timing of yellow change and red clearance intervals was formulated on the basis of a kinematic equation and its associated variable values. The recommended guideline encourages a uniform practical application that provides a framework that can be easily adopted into transportation agency practice.

In 2009, NCHRP Project 03–95 sought to develop a comprehensive, uniform set of recommended guidelines for determining safe and operationally efficient yellow change and red clearance (i.e., all-red) intervals at signalized intersections. Before guideline development, a comprehensive field study of driver behavior was performed to establish the three primary parameters found in a commonly used kinematic equation (the ITE equation): perception–reaction time, deceleration rate, and approach speed. To provide broad nationwide representation of driver characteristics, site conditions, and traffic signal timing practices, 83 study approaches were evenly selected from four regions of the United States: Southeast Michigan, Central Florida, Southern California, and Greater Washington, D.C. Driver behavior data were collected at each site with an elevated high-definition video camera. The resulting data set included 2,458 first-to-stop dilemma zone vehicles that were less than 5.5 s upstream of the intersection at the start of the yellow interval. The study's overall mean brake response time of 1.00 s was identical to ITE's recommended default value of 1.0 s for timing of the yellow interval, and the overall mean deceleration rate of 10.08 ft/s² was very close to ITE's recommended deceleration rate of 10 ft/s². The study determined that the 85th percentile speed of free-flowing through-moving vehicles approaching a signalized intersection may be accurately estimated by adding 7 miles per hour (mph) to the posted speed limit on the approach. Similarly, the 85th percentile approach speed for free-flowing left-turning vehicles may be accurately estimated by subtracting 5 mph from the posted speed limit.

The safety of crossing guards is important because it is their responsibility to make sure that schoolchildren cross the street safely and efficiently. Through discussions, it was found that crossing guards were very interested in using stop paddles with embedded lights to improve the conspicuity of the paddles. There was a concern, however, that some of the embedded light configurations might negatively affect a motorist's ability to recognize the three critical characteristics that define a stop sign: red background color, octagon shape, and white stop legend. A closed-course human factors study was conducted to assess the effectiveness of five stop paddles with embedded lights as compared with that of a standard, unlit stop paddle. On the basis of the findings from this study, the use of three stop paddle configurations is recommended to improve the conspicuity of crossing guards without negatively affecting a motorist's ability to recognize the three critical characteristics of a stop sign. These configurations are (a) a stop paddle containing flashing red lights arranged in an octagonal pattern at the eight corners of the paddle, (b) a stop paddle containing a series of steady-burn red lights around the border arranged such that the lights clearly convey the octagonal shape of the paddle, and (c) a stop paddle containing a series of flashing red lights around the border arranged such that the lights clearly convey the octagonal shape of the paddle. Because the Manual on Uniform Traffic Control Devices does not currently address the issue of steady-burn lights, agencies should contact FHWA before they use a stop paddle with a steady-burn light configuration.

High-speed signalized intersections present a grave hazard to the traveling public. An important component of this hazard is variation in the process of deciding whether to stop or proceed that occurs in the dilemma zone. To mitigate the dilemma zone problem, an innovative vehicle-specific detection, control, and warning system (DCWS) is being developed. To reduce dilemma zone conflicts, DCWS uses detection of individual vehicle speeds to provide unique, speed-specific dilemma zone protection for each vehicle as it approaches a signalized intersection. Although protecting all vehicles would be ideal, under some situations such protection is not possible. To account for these situations, DCWS also uses a vehicle-specific in-pavement system to give advance warning to drivers who are going to be trapped in the dilemma zone. With hardware-in-the-loop simulations, the new DCWS outperforms conventional dilemma zone warning systems in terms of the number of vehicles in the dilemma zone, the number and percentage of vehicles requiring warning, and the number of warning events per hour.

Two related evaluation studies that were conducted under NCHRP Project 17–35 are presented. The research objective was to use the most appropriate analytical methods and the best available data to build on previous research in developing crash modification factors (CMFs) for two treatments for reducing crashes related to traffic signal change intervals: modifying the change interval and installing dynamic signal warning flashers. Three evaluation methods were used as appropriate—the empirical Bayes before–after method, the comparison group before–after method, and cross-sectional multiple regression models. A secondary objective of using cross-sectional models for some evaluations was to examine the comparability of before–after and cross-sectional studies, a subject of topical interest in CMF development. There was a general safety benefit to installing dynamic signal warning flashers, with indications that crash reductions could be obtained overall and for several crash types, including injury, angle, and heavy-vehicle crashes. For the change-interval modification, the before–after study results showed significant reductions (at the 5% level) in total, injury, and rear-end crashes under various scenarios. For both treatments, the results from the cross-sectional analysis were relatively consistent with those from the before–after analysis.

This paper documents a research project that evaluated proposed new traffic signs to ensure that the signs were effective when human factors requirements were taken into consideration. The following sign messages were selected for symbol development and evaluation: combination horizontal alignment–advisory speed, congestion ahead, cross-street preferential lane warning, do not enter, do not pass, electric-vehicle charging station, fallen rocks, flagger ahead, low-shoulder warning, maximum width, no left turn ahead, railroad crossing on leg of roundabout, road narrows, survey crew, trolley crossing, uneven lanes, and winery. The team conducted 96 surveys to determine driver comprehension and legibility for each alternative sign design. The surveys were administered to the general driving public at the Turner–Fairbank Highway Research Center in McLean, Virginia. On the basis of driver input and the results of the comprehension and legibility testing, the research team provided recommendations on symbols that should be included in the next edition of the Manual on Uniform Traffic Control Devices.

With the use of a general classification regression model, this study investigated the causal relationship between time to train arrival (TTA) and crash frequency at highway–railroad grade crossings. In particular, a stratified structure in the explanatory variables was used to avoid the collinearity problem generally confronted in linear regression models. TTA is a good estimate of rail sight distance and time to collision, and it could be used to predict crash frequency at a grade crossing. A 14-year crash data set accompanied by crossing inventory data including TTAs was collected for the empirical study. Study results indicated that a negative relationship between TTAs and crash frequencies was generally found for all types of trains. Similar causal relationships were also found in various combinations of both crossing attributes and crash characteristics. Sensitivity analysis on the variable combinations was also conducted to investigate the key risk factors that might result in traffic collisions at grade crossings. Policy implications based on the empirical study are discussed, and future research directions are recommended.

Safety at railway level crossings (RLXs) is a worldwide issue that increasingly attracts the attention of relevant transport authorities, the rail industry, and the general public. The differences in the operation characteristics of varying types of warning devices, together with differences in crossing geometry, traffic, or train characteristics, leads to different driver behaviors at crossings. The aim of this study was to use traffic microsimulation modeling based on field video recording data to compare the safety performance of varying conventional RLX warning systems. The widely used microsimulation model VISSIM was modified to produce safety-related performance measures, namely, collision likelihood, delay, and queue length. The results showed that RLXs with an active warning system were safer than those with a passive sign by at least 17%. Integration of surrogate measures in conjunction with traffic simulation models determined which safety approach was more efficient for specified traffic and train volumes.

Driver behavior is influenced by environmental lighting conditions on roads; in the literature, many studies report a reduced night–day accident ratio following improvements to lighting on different types of roads, with the results classified by severity and type of accident. Few studies, however, report the influence of lighting conditions on driver speed. This study investigates the principal factors that influence driver speed on arterial roads in Turin, Italy. The aim of this study was to analyze driver speed under different daylight and nighttime lighting conditions. Six arterial roads were selected for observation and the measurement of speeds and illuminance on the pavement surface. The results showed that illuminance, in addition to factors such as lane position, lane width, and the relevant speed limit, should be considered a variable that can influence driver speed. The study used a regression equation to predict operating speeds (V₈₅) on urban arterials; the corresponding sensitivity analysis has made it possible to quantify the effects of the aforementioned variables on operating speed under different environmental lighting conditions.

This study evaluated high-beam usage on low-volume roadways in Texas. Most previous studies used a subjective approach to determine high-beam usage that involved a stationary observer or an observer driving around the study section. For more efficient data collection, this study used photometric readings as the methodology. This approach allowed for a quantitative measure of high-beam use, in that it did not rely on an observer's opinion as to whether the headlamps were on high or low beam. Preliminary data collection was conducted as part of the study to estimate the threshold high-beam illuminance value; these illuminance data showed moderate use of high beams on low-volume, rural two-lane highways, with 42% of the free-flow drivers using high beams. There was a wide variation in high-beam use between sites, so site characteristics were reviewed to identify potential reasons for this variation. The data were fit to a logistic regression model to estimate the impact of various site characteristics on a driver's decision to use high beams. The results showed that vehicle speed, paved surface width, number of access points within 1,000 ft of the site, and presence of horizontal curves within 1 mi of the site were statistically significant factors that contributed to the probability of high-beam use. The results also showed that higher vehicle speeds, narrow paved surfaces, presence of horizontal curves, and few or no access points led to a higher probability of high-beam use. These results are intuitive and support the observations made in the exploratory data analysis.

Driving is often considered a visually oriented task. This visual task is constrained when drivers drive at night. Visibility is reduced because visual cues available during the day are not present at night. This study attempted to develop a link between driver safety and the nighttime visual environment. This research required creating and integrating new technology, observing and collecting data, and developing a model framework called the Dynamic Driver Visual Model (DDVM). Conceptually, a DDVM is a system of rules, statistics, and expectations that can be used to define how a driver collects visual information from the environment. Several information sources were investigated, and several dependent variables were identified. Data were collected on how information from signage, objects, lighting, pavement markings, and other vehicles moderates a driver's visual search of the roadway environment. Several logistic regression analyses were performed on the collected data to identify common characteristics to be implemented in the DDVM. These variables included age, lighting, vehicle headlamps, several different objects, glance time, target luminance, and contrast information. The results suggest that a number of target and visibility elements have nonlinear effects on a driver's detection performance at a variety of detection distances. This paper discusses the implications of these findings and the initial framework of the DDVM. Future research and additional data requirements are also discussed.