A new software tool helps highway agencies tackle road departure crashes, reducing their consequences.

Each year, more than 16,000 people are killed and another 1 million are injured in run-off-the-road (ROR) vehicle crashes. This type of crash occurs when vehicles leave the travel lane, encroach onto the shoulder and beyond, or hit one or more natural or man-made objects, such as utility poles, bridge walls, embankments, guardrails, parked vehicles, or trees. In recent years, ROR crashes have been on the rise, making up 33 percent of total crashes in 1995 and 38 percent in 2000.

According to the Federal Highway Administration's (FHWA) 1998 National Strategic Plan, roadside crashes cost society $80 billion each year. This is more than three times the annual amount spent by Federal, State, and local government agencies to maintain and operate our Nation's roads. According to the National Cooperative Highway Research Program (NCHRP), "The annual societal cost of roadside accidents is enough to purchase 350 gallons of gasoline for every registered vehicle in the United States."

Numerous government agencies and organizations are working to combat the ROR problem. The NCHRP, in its Strategic Plan for Improving Roadside Safety, stated its vision: "A highway system where drivers rarely leave the road; but when they do, the vehicle and roadside work together to protect vehicle occupants and pedestrians from serious harm."

By 2007, FHWA would like to reduce fatalities involving roadway departure crashes by 10 percent—one of the goals of the agency's Vital Few strategy. (Road departure crashes include both ROR and head-on crashes and are combined into one category because similar countermeasures can be used to tackle both problems.) To meet this objective, FHWA's approach includes research, working cooperatively with other organizations, and developing and implementing strategies such as a new computer analysis tool.

Researching Road Departure Safety

Addressing ROR crashes presents significant challenges because of the vastness of the road network and the randomness of these crashes. Two out of three ROR fatal crashes occur in rural areas, 80 percent occur on dry pavement, and 60 percent occur during dark or reduced light conditions.

To address the problem of road departure safety (RDS), FHWA is developing an Interactive Highway Safety Design Model, which is a set of evaluation tools for assessing the safety effects of specific geometric design decisions. FHWA also is investigating the crashworthiness of roadside and roadway features, as well as the means to ensure minimum retroreflectivity for signs and pavement markings to improve nighttime visibility.

Working Together

FHWA is not fighting the battle against road departure crashes alone, as the agency is working cooperatively with other organizations on several projects. In conjunction with the American Association of State Highway and Transportation Officials (AASHTO), the National Highway Traffic Safety Administration (NHTSA), and the Transportation Research Board (TRB), FHWA assisted with the development and implementation of AASHTO's Strategic Highway Safety Plan. The plan includes 22 key emphasis areas that affect highway safety and potentially could help save lives. Five of the 22 areas specifically address road departure crash concerns—hazardous trees, ROR crashes on two-lane rural roads, head-on crashes on two-lane rural roads, utility poles, and horizontal curves.

FHWA also is helping State and local agencies increase public awareness of their RDS programs and build support for those efforts. Through a coalition of government, industry, institutional, and civic partners, including the National Safety Council, the Roadway Safety Foundation, the American Traffic Safety Services Association, and AASHTO, FHWA is helping disseminate information about State and local agencies' roadside safety programs through the use of Web sites and knowledge management exchanges.

Researchers need the ability to model where, when, and how road departure crashes occur to understand the problem fully and to develop effective countermeasures. To provide the tools and information that researchers need to model these crashes, FHWA and AASHTO will work together to develop a road departure knowledge base that will contain existing crash data and information gathered from focused RDS studies. To supplement the RDS knowledge base, FHWA is conducting comprehensive rollover research to determine the cause of rollover crashes, investigate effective countermeasures, and develop guidance for minimizing rollovers.

Let's Get Grooving Run-off-the-road crashes cause one-third of all traffic fatalities. One of the main causes of ROR crashes is driver fatigue, which is often compounded when drivers simply drive too fast. Alcohol and other drugs can contribute to both fatigue and speed, but most often the problem is drowsy drivers who think they can "make it home" and then become run-off-the-road crash statistics. Statistics from the 2001 Fatality Analysis Reporting System (FARS) show that approximately 31 percent of all fatal crashes were categorized as single-vehicle run-off-the-road crashes. Shoulder rumble strips are one way to address this significant safety problem.  Rumble strips are raised or grooved patterns constructed on the roadway's shoulder. Vehicle tires passing over them produce a rumbling sound and cause the vehicle to vibrate. The noise and vibration produced by the strips are effective alarms for drivers who have drifted from their travel lane onto the shoulder. They are used primarily on expressways, interstate highways, and parkways, although some States are beginning to install them on two-lane rural roads that have high numbers of single-vehicle crashes. How effective are shoulder rumble strips as a safety enhancement? Several studies indicate that they can reduce the overall rate of ROR crashes by 15 to 70 percent, which would lead also to a reduction in the number of injuries and fatalities. Rumble strips have their drawbacks, including bicyclists' concerns about safety. Taking into account the combined weight of a bicycle and bicyclist compared to a vehicle, the vibration produced when a bicycle passes over the shoulder rumble strip can be considerable. Although deeper and wider shoulder rumble strips have been shown to be more effective for warning drivers, deeper shoulder rumble strips can make it more difficult for bicyclists to retain control of their bicycle while crossing the strips, even at low speeds. Because of these concerns, FHWA's guidance on placement of shoulder rumble strips is a minimum of 1.2 meters (4 feet) of paved shoulder to the right of the edgeline, or 0.3 meters (1 foot) to the right of an edgeline on narrow shoulders where shoulder rumble strips are at the outside edge of the paved shoulder. FHWA has developed a Web site to address the crucial role of shoulder rumble strips. The site includes FHWA's Technical Advisory on Roadway Shoulder Rumble Strips and a synthesis of rumble strip information. The technical advisory provides FHWA's guidance on where and when rumble strips should be used. The synthesis is a review of the current practices of State departments of transportation, a review of recent shoulder rumble strip studies, and a review of common practices. The Web site can be found at http://safety.fhwa.dot.gov/roadway_dept/rumble/synthesis/pro_res_rumble_library.htm. AJ Nedzesky and Richard Powers

Road Departure Safety Strategies

Federal, State, and local agencies can choose from an array of strategies to help reduce the number of road departure crashes. Regardless of which strategies an agency selects, the objective should be to keep drivers and vehicles on the road and to minimize the severity and likelihood of crashes when vehicles do leave the road. In the perfect world, the first objective would be met by designing roads that help keep vehicles in the travel lanes. However, even proper roadway design cannot prevent all road departure crashes.

Inadvertent roadside encroachments occur for a variety of reasons, including a vehicle swerving to avoid another vehicle or object, driver fatigue, weather-related hazardous road conditions, or driving too fast for conditions. Roadway improvements that tend to keep vehicles on the road include rumble strips, better geometric design, increased skid-resistant roadway surfaces, more durable pavement markings, and more visible roadside signs.

When vehicles do leave the road, strategies are needed that minimize the likelihood or severity of the potential crash. The probability of a crash occurring depends on roadside features such as the presence and location of fixed objects, shoulder drop-offs, side slopes, ditches, and trees. The probability of a crash is minimized if the roadside is fairly flat, without objects, and the soil can support vehicle tires. If a crash does occur, making the roadside hardware more forgiving and modifying side slopes to prevent rollovers can minimize the severity.

Specific Strategies

As mentioned, numerous strategies are available to reduce the number of road departure crashes and minimize the consequences. First, highway agencies should upgrade obsolete guardrail terminals such as blunt or turned-down ends. Guardrails are responsible for nearly 1,200 fatal crashes each year—a third of those occurring at the terminals.

A second way to minimize crash consequences is developing and implementing median barrier warrants and using median barriers such as cable guardrail.

Third, highway agencies should be encouraged to implement the new guidelines that are being developed for accommodating heavyweight mailboxes. An increasing number of people are purchasing heavy-duty mailboxes or installing secure mailboxes to prevent theft and vandalism. These mailboxes pose a threat to vehicles and drivers when crashes occur.

A fourth strategy is to develop and promote a sustainable safe tree and urban gateway policy—landscaping and monumental signs at jurisdiction boundaries—and to teach "safe landscaping" courses that emphasize reducing the number of trees and utility poles near the roadway. Road departure crashes with trees account for nearly 25 percent of all fixed-object fatal crashes.

Fifth, the consequences of roadside departures can be minimized by developing programs that reduce the number and severity of horizontal curve crashes, which account for one-fourth of all highway fatalities.

Finally, resurfacing should include safety upgrading with paved shoulders and the installation of a 45-degree-angle asphalt fillet (safety edge) along each side of the paved edge. This measure is especially important along two-lane rural roads with narrow lanes and unpaved shoulders with pavement edge drop-offs resulting from roadway resurfacing.

Computer Analysis

To address some of the issues mentioned above, FHWA is using a new computer analysis tool called Finite Element Analysis (FEA) that will assist highway agencies with decisionmaking to reduce the severity of impacts into roadside features. FEA is a computerized tool that simulates crashes using computational techniques in structural mechanics. Most importantly, FEA can provide guidance for roadside design decisions that previously would not have been cost-effective or feasible.

Finite element modeling involves producing a Computer Aided Design (CAD) representation (called a "mesh") of the complete geometry of a physical structure by dividing it into many small elements to model all details. For crash modeling, meshes are created of vehicles and roadway features. The vehicle's mechanical properties are defined and associated with groups of elements representing various structural parts. Geometric data for the vehicle and any roadside features are obtained by measuring the actual objects.

Once the meshes are created, a computer program such as LS-DYNA, the FEA program used by FHWA, simulates a crash using physics-based equations to determine the vehicle or roadside feature's response to various loading and restraining conditions. LS-DYNA simulates an impact, such as a vehicle crashing into a portable barrier, by looking at the impact conditions and the impacting object's physical properties, and then calculating the forces and material reactions by individual time step for each element in the vehicle's mesh. The program then provides detailed information about the crash's impacts on the entire vehicle.

The level of detail and accuracy associated with FEA increases as the number of elements representing a single structure increases. A vehicle represented by 15,000 elements used in an FEA simulation will produce much less detailed results than if the same vehicle were represented by 100,000 elements.

Increasing the number of elements also increases the time needed for FEA programs to model crashes. For example, a parallel processor or super computer would take several hours to model one-sixth of a second of an accident involving a vehicle crashing into a piece of roadside hardware, if each structure were made up of only 30,000 elements. As the number of elements grows and finer structural details are included, the computational time increases significantly. Today, it is not uncommon for models to take 24 to 48 hours to compute crash simulations based on vehicles comprised of 400,000 elements, crashing into roadside hardware comprised of 50,000 elements or more.

Using FEA

Highway agencies often are tasked with evaluating the safety of roadside features. In many cases, following design standards such as those issued by AASHTO provides some guidance for typical road and roadside conditions. Crash tests of roadside features or hardware on flat and level ground also are useful and can provide a great deal of information.

However crash tests are not always cost- or time-effective when researchers want to evaluate the impacts of various crash scenarios on different vehicles and roadway features. For example, several crash tests would be needed to test the many alternatives used for addressing road discontinuities such as ditches, transitions between embankments and slopes, grading around flared terminals, and pavement drop-offs. FEA is more effective for selecting the best alternative in situations such as this, as well as for choosing among mitigation measures for road departure crashes.

Determining the appropriate height of highway signs provides one example of how FEA has been used. Research originally showed that U-post signs at a height of 1.5 meters (5 feet) would hit the windshield of small cars, while 2-meter (7-foot) signs would not. This information led transportation professionals to propose that the Manual on Uniform Traffic Control Devices' policy of posting rural signs at 1.5 meters be changed.

However, as the composition of the vehicle fleet changed, it was necessary to reevaluate whether the 2-meter signs are still compatible with today's vehicles. FEA was used to make this determination. The results showed that small cars were safer when signs were posted at 2 meters, and pickup trucks were safe at both the 1.5- and 2-meter heights. However, minivans were predicted to have windshield impact at 2 meters. The results of one FEA simulation were validated by a crash test and based on these results, the sign height policy remains unchanged at 1.5 meters.

Current projects using FEA include evaluations of cable guardrail penetration and the performance of the backside of median W-beam rails. FEA also is being used to develop:

• Performance limits for secure and vandal-proof mailboxes and for elevated medians when used for the placement of trees and other landscaping
• Usage guidance, size requirements, and performance limits for portable concrete barriers
• Height tolerances for W-beam guardrails
• Performance limits for curb-guardrail combinations

Note that FEA is a predictor tool and does not replace crash testing for initial acceptance. However, when validated, FEA simulations could provide more data on vehicle and roadside hardware performance than available from any crash test. In fact, the FEA simulation results could produce previously unavailable data on the structural behaviors (deformations, stresses, and strains) for nearly all vehicles or roadside features.

Enhancements in FEA

Use of FEA to supplement road safety designs started in the early 1990s, but its use has been limited due to a lack of specific vehicle and roadside feature models. Researchers already have developed models for 10 vehicle types and 7 roadside hardware features, with more to be available soon.

Presently FEA is better able to compare design alternatives than absolute crash test results. As researchers further refine the models, they will be better able to identify the boundaries of their accuracy. For example, for different portable concrete barrier designs, they have more confidence in predicting the strength of the connections than they have in predicting whether rollover will occur.

In the future, FEA could be used for developing more performance limits of roadside features or combinations of features, creating surrogate tests, or identifying and collecting information about rare failures in roadside safety features.

To make the tool more useful in the future, efforts are underway to establish guidance on the quality of finite element models, develop FEA analytic procedures and methodologies, and increase FEA analyst training. Increasing the knowledge of computer code limitations also is needed, as well as expanding the availability of adequate computer hardware.

Safer Roadsides for Tomorrow

As highway agencies increase their use of strategies such as rumble strips and stripes, fewer vehicles will depart from the road inadvertently. However, roadway departures are inevitable. When they do occur, the consequences must be minimized by reducing the severity of the impact. FEA can help overcome difficulties in developing guidance in the use of roadside safety countermeasures and enables highway agencies to make roadside safety decisions cheaper, faster, and better.

Harry W. Taylor is the road departure safety team leader in the Office of Safety Design. He has been involved in highway and roadside safety work for more than 25 years. He has participated on numerous NCHRP panels and industry-government roadside safety groups. Taylor is also one of two U.S. observers to the European Committee on Standardization (CEN), Technical Committee 226, Working Group 1, Road Equipment-Safety Barriers. He has a bachelor's degree in civil engineering from Tennessee State University and a master's in engineering administration from The George Washington University.

Leonard Meczkowski is a highway safety specialist and roadside team leader in FHWA's Office of Safety R&D. He manages the FHWA/NHTSA National Crash Analysis Center and has been involved in roadside safety work for more than 28 years. He is a graduate of Henry Ford College.

To learn more about FHWA's roadside safety program, visit www.fhwa.dot.gov/research/topics/safety/roadwaydeparture/. For more information about FEA, contact Leonard Meczkowski at 202-493-3317.