|Back to Resources|
Intersection safety is and has been a major program at the Federal Highway Administration (FHWA). A highway intersection is defined in this context to be the at-grade junction of two or more public roads. Within intersections, vehicle-vehicle and vehicle-pedestrian conflicts occur as drivers, bicyclists, and pedestrians need to cross the path of other vehicles. Not surprisingly then, the intersection, whether or not it is under traffic signal control, can be a hazardous location as evidenced by various motor vehicle crash statistics.
Using the year 2002 crash database maintained, as part of the National Accident Sampling System, there were an estimated 2.7 million motor vehicle crashes at intersections in the United States in 2002. Approximately 1.4 million crashes were reported at unsignalized intersections and 1.3 million were crashes reported at signalized intersections. An estimated 925,000 people were injured in the crashes at intersections. Approximately 445,000 people were injured in crashes at signalized intersections and nearly 480,000 were injured in crashes at unsignalized intersections. In terms of fatalities based on the Fatality Analysis Reporting System, a total of 9,117 people died in 2004 as a direct result of crashes for which the relation to junction was classified as at an intersection (non-interchange) or intersection-related (non-interchange).
In May 2002, the FHWA and the American Association of State Highway and Transportation Officials (AASHTO) sponsored a scanning study of signalized intersection safety in Europe. With the goal of improving signalized intersection safety, the focus of the scanning study was on innovative signalized intersection safety practices in Europe. In December 2003, the FHWA published the final report on the study, which was entitled Signalized Intersection Safety in Europe as part of FHWA's International Technology Exchange Program. Given the relative success of that European scan and feedback on the report, FHWA decided that a scanning study should be conducted of select areas in the United States. It is hoped that a Domestic Scan will also produce tangible benefits through identifying and making highway agencies aware of innovative treatments and practices that have been successfully implemented in the United States.
The Domestic Intersection Safety Scan was conducted in February 2005. The goal of the scan was to reduce fatalities, personal injuries and crashes at intersections in the United States by promoting innovative intersection treatments and comprehensive intersection safety processes that have been implemented in this country.
One of the primary objectives was to identify and document selected innovative intersection treatments that have been implemented and have been demonstrated to or have the potential to improve safety. Another objective was to identify and document selected comprehensive safety processes and procedures that have been implemented by transportation agencies specifically to improve intersection safety.
The range of treatments considered within the scan included treatments that addressed and/or contained the following elements:
During the scan, a secondary objective was to gain knowledge about and document the processes and procedures that were employed to gain agency management's approval for successful implementation and deployment. To the extent that the treatments had been evaluated by local agencies, the scan endeavored to gain knowledge about the safety effectiveness of these treatments and comprehensive approaches to intersection safety.
Because of limited resources, the scan team could not go to all areas of the United States to cover all relevant topics. A vetting process was conducted to identify areas where innovative intersection treatments and comprehensive safety processes were implemented and the local agencies were recognized for their prominence in selected areas and a willingness to participate and share their knowledge, time, and experiences. The following identifies the agencies and organizations that participated in the scan and their innovative intersection treatments and comprehensive safety processes:
Like many of the states, counties, cities and municipalities throughout the United States, the host agencies that participated in this scan are endeavoring to improve safety at intersections. The treatments that they have developed and implemented are not applicable to all intersections and may not be appropriate for a given municipality or state.
Several host agencies agreed that the first step toward achieving significant improvements in intersection safety is to create a culture of safety within the organization. These organizations found that by assigning a greater prominence to safety in transportation investment decisions, they were able to produce significant reductions in crashes. Before this could happen, it was understood that the agencies had to raise the awareness and importance of highway safety throughout all branches of the state, county, city, and municipal government transportation departments. This required the development and implementation of processes and procedures to monitor the performance of the highway system in measurable safety criteria, including crash frequency, rates, and severity.
Safety management that is truly performance-based was judged to be the cornerstone. The greatest gains were experienced by those agencies that had established formal numerical goals and measurable objectives with respect to crash experience. Finally, several of the host agencies also pointed to public-private partnerships as a means of improving intersection safety. The project completed by the cities of Detroit and Grand Rapids, in association with AAA Michigan, was cited as a notable case study. Since AAA Michigan is somewhat unique in that it is an insurance provider as well, other business models could be applied to better fit the constraints and opportunities that exist throughout the United States.
Repeatedly, host agencies indicated that it was not possible to do a reasonable job in intersection safety unless accurate crash data was matched to the correct intersection. Similarly, the host agencies voiced many concerns attributable to highly suspect crash data. All levels of government must assume a commitment to improving the quality of crash data, as well as supporting intersection inventory data and traffic data. Without a set of clearly defined numerical goals and established performance standards, operating agencies will continue to wait excessively long periods until crash data becomes available for their use.
Most host agencies also indicated that better access to crash data is needed to further enhance intersection safety. Specifically, they desired to have quickly-generated spatial data displays. Agencies with access to tools that allow generating Geographic Information Systems (GIS) pin maps and other displays, such as those that can be generated from the SEMCOG Web site, felt empowered.
Several of the host agencies also discussed a need for a flow of safety-related information from the state's central agency. The scan revealed that there are multiple benefits to the two-way exchange of crash and intersection-related data. Certainly, many benefits accrue when data can be transported up from local police departments and the state patrol to the appropriate headquarters agency tasked with the responsibility for the central crash records system. However, the benefits are also large when the data is reduced, subjected to quality control checks and summarized in meaningful formats, and returned to the police departments and transportation agencies at the local government level.
A few of the host agencies conducted rigorous before and after evaluations of the effects of these implemented treatments on crash experience. Therefore, there is still knowledge to be learned about many of the treatments cited in this report. Limited sample size and limited post-treatment durations restrict the evaluator's ability to generate strong conclusions on the effectiveness of these treatments.
In addition, there is a need to develop and maintain an accurate knowledge base of the effects of projects, including those with multiple treatments, on crash experience. Safety effectiveness estimates are especially needed for flashing yellow left arrow signal indications, pedestrian detection systems that seek to extend pedestrian clearance intervals, and treatments that delay the onset of the yellow interval, among the treatments encountered during the scan.
Within the area of traffic control devices at intersections, many of the host agencies implemented innovative treatments, including street name signs with larger lettering in Clearview font at signalized intersections and advance street name signs that were placed at locations on the major approaches upstream of the intersection. There were also numerous pedestrian treatments, including pedestrian countdown devices, more pronounced crosswalk markings, audible pedestrian signal heads, and pedestrian push buttons and signs designed for mobility-challenged pedestrians. Some have implemented activated, in-pavement lights for crosswalks and activated pedestrian crossing warning devices-systems that alert drivers of possible conflicts.
With respect to pavement markings, several host agencies made innovative use of dashed markings, which are frequently called "cat tracks" or "puppy tracks" text on pavement surfaces at locations where supplemental directional information is needed, and messages (e.g., "LOOK LEFT") in the pavement where there is a greater need to communicate to pedestrians, such as a roundabout.
Some agencies installed internally illuminated traffic sign boxes, which are continuously lit at night that featured permanent regulatory restrictions, such as "NO LEFT TURN" and "STOP." Other agencies installed internally illuminated "PEDESTRIAN CROSSING" signs at mid-block crossings that are illuminated in response to actuations of pedestrian push buttons. Fiber-optic and other dynamic regulatory signs (e.g., "NO TURN ON RED" to communicate time-dependent regulations) were effectively used by some host agencies.
Virtually every host agency identified traffic operations strategies and techniques that most would consider conventional practices. These included yellow and red clearance intervals for phases at signalized intersections that meet national guidelines/practices, the provision of additional crossing time for older pedestrians and at intersections where conditions warrant (e.g., high numbers of elderly pedestrians and/or school children), and protected left-turn phases that can be called back into service during the same cycle under certain conditions.
Many of the host agencies also described their experiences with innovative practices. In addition to the "Dallas Phase" sequence for left-turn movements at intersections operating with lead-lag left turn phasing, the city of Dallas cited an innovative treatment that allows for longer pedestrian walk and pedestrian clearance intervals to be subsequently provided in response to continuously depressing the pedestrian push button for five seconds or more. The cities of Portland and Richardson have experimented with systems that delay the onset of the yellow interval or extend red clearance intervals, respectively. While it is common practice to vary the duration of green intervals in response to congestion conditions, it is hoped that in the future systems, processes, practices and/or procedures can be devised that would allow for the dynamic variation of yellow, red clearance, pedestrian walk and pedestrian clearance intervals in response to monitored conditions to reduce safety risks.
Other innovative treatments implemented included delaying the onset of the yellow interval based on detection of vehicles beyond the stop line at wide intersections, implementing longer pedestrian walk and pedestrian clearance intervals at different times of the day when students are present at intersections near schools, time-of-day phasing in which left-turn phasing sequencing can be varied by time of day and day of week, and flashing yellow left turn arrows.
Innovative, non-traditional geometric design treatments, which were implemented by the host agencies, included the Michigan Indirect Left Turn treatment, the New Jersey "jug-handle" treatment, roundabouts, mini-roundabouts, and the Michigan "Loon," which facilitate U-turns by large trucks at intersection sites with narrow medians and less than three opposing travel lanes. Several host agencies also implemented road diet projects that converted four-lane cross sections to three-lane cross sections. Median island treatments were constructed on several intersection approaches to limit the effects of nearby driveways and other access by eliminating crossing maneuvers from minor access points. Other treatments include intersection bulb-outs and offsetting left turn lanes to improve sight distance at intersections with opposing left turn lanes and permitted left turn signal phasing.
Several of the host agencies had installed unique crosswalks, which included brick crosswalks, blue bike lanes, raised crosswalks, and novel intersection designs, such as raised intersections that are similar to speed tables. At a few downtown intersections where some of these treatments were constructed, the corner curbing was concurrently removed to make the corner landing flush with the roadway. Although these geometric treatments are clearly innovative, there is a healthy debate on their appropriateness by road, functional classification, area context, vehicle speeds, and volume.
Although enforcement and driver education were not the focus of this scan, several notable programs were identified by host cities. Enforcement treatments implemented by the host agencies pertained to the vigilant enforcement of unsafe driving behavior at intersections. These included so-called "rat lights," which assist police agencies to identify drivers that violate red signal indications at intersections, and photo enforcement systems.
Other enforcement programs featured enforcement target maps, which have been developed by traffic engineering agencies to pinpoint the clustering of selected crash types for given "targets," such as red-light running, speeding, and aggressive driving, among others. Driver education treatments for intersection safety included multi-jurisdictional education campaigns aimed at red light running and widely distributed brochures in Spanish that explain traffic control devices at intersections.
Other innovative education programs included a mobile truck simulator for truck driver training and portable, radar-based dynamic speed signs. Innovative traffic control or geometric design treatments should not be implemented without advance public information. Wherever new treatments, especially those that are non-intuitive, are to be implemented, consideration should be given to developing and conducting a comprehensive public education program prior to deployment and updating driver training materials to ensure that the message is communicated to new drivers.
Readers of this scan report are encouraged to contact the individuals in Appendix B to learn more about the treatment and processes that are described. It is hoped that better and more effective intersection safety treatments can be developed and implemented by others as a result of this search for innovative ideas.
Some of the traffic control devices or applications described in this report are not in compliance with the Manual on Uniform Traffic Control Devices (MUTCD) and are considered experimental. Any jurisdiction wishing to use a non-compliant device or application on a road open to public travel must request and receive approval from the Federal Highway Administration for experimentation. Please refer to Section 1A.10 of the MUTCD (http://mutcd.fhwa.dot.gov) for procedures regarding "experimentation."
Previous | Table of Contents | Next
back to top