COMMITTEE ON ROADSIDE SAFETY FIGURES

John F. Carney III, Chairman

D.W. "Bill" Dearasaugh, Jr., TRB Staff Representative

July 27-30, 1997

This workshop was intended to promote a dialogue on future directions in roadside safety among government officials, industry representatives, researchers, academics, and others. The strategies for improving roadside safety evolving from the efforts under NCHRP Project 17-13, Strategic Plan for Improving Roadside Safety, were presented. The need for improved standards for roadside safety features was the focus of the workshop. Workshop participants had the opportunity to raise questions about current standards, assess their relevancy, and propose improvements. The results of this workshop will provide the basis for considering research efforts to be undertaken under NCHRP Project 22-14, Improvement of the Procedures for the Safety-Performance Evaluation of Roadside Features.

WELCOME
Daniel W. Dearasaugh
Transportation Research Board

FORMULATING A STRATEGIC PLAN FOR IMPROVING ROADSIDE SAFETY
John F. Carney, III
Worcester Polytech Institute

TRANSLATING STRATEGIES INTO ACTIONS
Richard G. McGinnis
Bucknell University

IMPLEMENTATION AND RESEARCH NEEDS
Kenneth S. Opiela
NCHRP, Transportation Research Board

UPDATING PROCEDURES FOR THE CRASHWORTHINESS EVALUATION OF HIGHWAY SAFETY FEATURES
Charles D. Sanders
Illinois DOT

WORK PLAN FOR UPDATING PROCEDURES FOR SAFETY PERFORMANCE EVALUATION
King Mak
Texas Transportation Institute

DEFINING AND MEASURING RELEVANCY OF EVALUATION PROCEDURES
Lindsay I. Griffin, III
Texas Transportation Institute

REVIEW OF TEST MATRICES AND CONDITIONS
Roger P. Bligh
Texas Transportation Institute

OTHER SAFETY PERFORMANCE EVALUATION CRITERIA
Maurice E. Bronstad
Dynatech Engineering, Inc.

EFFICACY OF ALTERNATE EVALUATION TECHNIQUES
Dean L. Sicking
University of Nebraska

EVALUATION OF ROADSIDE FEATURES TO ACCOMMODATE VANS, MINI-VANS, PICKUP TRUCKS AND 4-WHEEL DRIVE VEHICLES
Hayes E. Ross, Jr.
Texas Transportation Institute

FHWA PERSPECTIVE ON FULL-SCALE CRASH TESTING
James H. Hatton, Jr.
Federal Highway Administration

FHWA PERSPECTIVE ON FULL-SCALE CRASH TESTING
Richard D. Powers
Federal Highway Administration

TESTING AGENCY VIEWS
John Rohde
University of Nebraska

TESTING AGENCY VIEWS
Wanda L. Menges
Texas Transportation Institute

HARDWARE INDUSTRY PERSPECTIVE
Owen S. Denman
Energy Absorption Systems

HARDWARE INDUSTRY PERSPECTIVE
Donald Johnson
Trinity/Syro Industries

HARDWARE INDUSTRY PERSPECTIVE
Dean C. Alberson
Safety Quest, Inc.

SUMMARY OF BREAKOUT GROUP DISCUSSIONS
Group A- Relevancy of Evaluation Procedures (Moderator- Lindsay I. Griffin, III)
Group B- Test Matrices and Conditions (Moderator- Maurice E. Bronstad)
Group C- Full-Scale Crash Testing (Moderator- Owen S. Denman)
Group D- Evaluation Criteria (Moderator- Richard D. Powers)
Group E- Processes for Certification (Moderator- Don J. Gripne)

APPENDIX A-- PARTICIPANT LIST

I would like to welcome everyone on behalf of the Transportation Research Board. Your attendance and participation are greatly appreciated.

People talk about seeking excellence and striving for improvements all the time. Whenever I hear those words, I think of how difficult they are to achieve. The TRB Committee on Roadside Safety Features (A2A04) is unique because it has been effective in meeting these goals. Continual improvement has been noted in increased attendance at meetings with a high level of interest. You are truly on the cutting edge of technology, have contributed to improved and increased funding for NCHRP projects, have strong leadership, and benefit from active subcommittees.

You may know that every third year each committee at TRB submits a triennial self evaluation. Dr. John Carney, the committee chair, prepared a self-evaluation and circulated it through the committee for review. When it was submitted to the TRB Group on Design and Construction of Transportation Facilities, A2A04 was the only committee at TRB that received a perfect rating on their triennial evaluation. This reflects the excellence of the committee.

This workshop deals with future directions in roadside safety. The results of NCHRP Project 17-13, Strategic Plan for Improving Roadside Safety, provide many suggestions for future directions. This project is winding down and it is providing some very useful information to TRB Committee A2A04, researchers, and highway safety professionals. One example is the new NCHRP Project 22-14, Improvement of the Procedures for the Safety-Performance Evaluation of Roadside Features. This project addresses the need for updated and improved procedures for evaluating the crashworthiness of roadside hardware given increased concerns about safety. This contract was awarded to the Texas Transportation Institute and work on the project has just begun.

This is a very appropriate time in the life of this project to seek input that will assure procedures that will serve future needs. It is hoped that this workshop will provide a lot of useful input into this venture, which will eventually result in changes, improvements and modifications in NCHRP Report 350, Recommended Procedures for the Safety Performance Evaluation of Highway Features. It is important to note the international interest in this subject. Testimony to this fact is the many guests at this workshop from other parts of the world.

A review of the strategic planning efforts for roadside safety will be presented in three parts. First, I will summarize the process undertaken over the last two years that led to the strategic plan. The five missions that are the basis of the plan will then be described. Many of you were present at the Park City workshop last summer during which this strategic plan was subjected to a thorough review. Efforts have continued to build on the goals, objectives, and actions based upon the feedback gathered at the meetings. Second, Dick McGinnis will discuss in more detail the objectives, and actions believed necessary to achieve the goals. Last, Ken Opiela will discuss the research needs and implementation activities that have been derived from the plan.

The roadside safety problem is a very serious one. We have a million highway crashes associated with the roadside every year in the United States. It accounts for one third of the highway fatalities, results in societal costs exceeding $60 billion, and clearly is a major cause on injury and death. It is a very complicated problem, because there are so many variables or external influences. These include:

• Changing vehicle fleet (size/design/weight)

• Increasing traffic volumes

• Aggressive driver behavior

• Environmental concerns

• Limited resources/budgets

• Right-of-way constraints

• Higher traffic speeds

• Infrastructure deterioration

• Older driving population

• Aesthetic considerations

In the last few years there has been increased interaction with General Motors to provide dialogue with the vehicle manufacturing sector relative to roadside safety. The vehicle fleet keeps changing and there is a need to work towards improved compatibility of vehicles and roadside features. Traffic demands keep going up and driver behaviors are changing. It is interesting to note that a couple of weeks ago Congress had a hearing on driver behavior and the director of NHTSA coined the phrase, "road rage." There have been a series of articles in the newspapers in the last few weeks talking about mild mannered people who get behind the wheel of a car and turn into monsters. This is a serious problem that can have an impact on roadside safety. Environmental concerns, higher traffic speeds, an aging driver population, and the other issues make it hard to get our arms around the roadside safety issue.

The U.S. has done very well in highway safety when you consider that the number of vehicle miles have increased exponentially and yet fatalities are down in the low 40,000s instead of 59,000, like they were some 20 years ago. But 42,000 people are dying a year in this country, which is totally unacceptable. To address the roadside safety problem it is necessary to consider all aspects - the roadway, the driver, and the vehicle. This was the motivation for starting this strategic planning project.

The first meeting on this project took place here exactly three years ago. The purpose of the meeting was to identify the issues. Transportation Research Circular 435 Roadside Safety Issues was the result of that meeting. The logistics for a project were discussed after the Task Force 13 meeting in Charleston. The first NCHRP 17-13 panel meeting took place in May of 1995 where a decision was made to solicit additional issues and employ a facilitator to develop the strategic plan. At the 1995 summer meeting of this committee at the National Academy's facility in Irvine, California, members of the committee discussed issues further and developed research needs in roadside safety. A facilitator led the panel through a lot of nitty gritty work to lay the groundwork for the strategic plan at that meeting. Then, in January of 1996, at the TRB Annual Meeting, a briefing for top officials was held. The problem was the blizzard, and if you didn't come into Washington on the Saturday, you didn't get there. This was a disappointment, because the briefing didn't get the attendance that had been expected. In May, back in Washington, the panel tried to draw outsiders into the discussions-state police, advocacy groups, public works officials, and others- to get their perspectives on the problems and approaches to solve the roadside safety problem. In Park City, Utah, the 1996 A2A04 Summer Meeting was devoted to refining the goals and objectives of the strategic plan. As a result of these efforts a comprehensive strategic plan has been completed and efforts are underway to document it.

The panel came to consensus on the vision for the strategic plan-a highway system where people do not pay with their lives when vehicles inadvertently leave the roadway. In this system, drivers rarely leave the road. When they do, the vehicle and the roadside work together to minimize harm.

For those of you who have never gone through the trauma of developing a strategic plan- and it is traumatic- these are the elements around which the plan is structured:

• Mission-- broad functional area needing attention.

• Goal-- a specific outcome expected to result from accomplishing the mission.

• Objective-- an aspect of a goal which can be measured.

• Action-- direct activities that can be undertaken to meet an objective.

• Need-- Knowledge or information necessary for an action to be undertaken.

Mission 1: Increase the awareness of roadside safety and support for it.

• Goal 1-- A coalition of governmental, industrial, institutional and civic partners that will work toward the improvement of roadside safety.

• Goal 2-- A heightened awareness of the importance of roadside safety by the public including citizens, decision makers, special interest groups, and others. (Environmentalist utility companies, MPO’s and grassroots organizations).

• Goal 3-- Increased emphasis on roadside safety by partners and stakeholders and better communications between them.

• Goal 4-- Sufficient financial resources to address roadside safety needs at the federal, state, and local levels.

• Goal 5-- On-going programs to disseminate roadside safety information to practitioners, researchers, decision makers, and the public.

• Goal 6-- A roadside safety component in all DOT safety management systems.

• Goal 7-- On-going process for updating the roadside safety strategic plan for prioritizing research needs.

• Goal 1-- Improved roadside and roadway inventory systems based on a common location referencing system.

• Goal 2-- Comprehensive roadside safety information resources including crash data, in-service evaluations, funding sources, research results, training programs, tort claims, highway inventories, traffic data, & vehicle sensor data.

• Goal 3-- State-of-the-art methodologies and tools for analyses of crashes and simulations of roadside accidents and crash tests.

• Goal 4-- On-going programs to conduct safety analyses and identify hazardous roadside locations.

• Goal 1-- Improved highway designs that reduce the probability of vehicles leaving the roadway.

• Goal 2-- Improved traffic operating environment that reduced the occurrences of events contributing to loss of vehicle control and roadside encroachment.

• Goal 3-- Sufficient maintenance of highways and vehicles to reduce the probability of loss of vehicle control.

• Goal 4-- Improve vehicle-based systems to help keep drivers on the road.

• Goal 5-- Improved driver performance and behavior.

• Goal 1-- Improved roadway geometrics and roadside designs that reduce the probability of overturns for the variety of vehicles using the road.

• Goal 2-- Improved vehicle designs that increase vehicle stability in run-off-the-road situations.

• Goal 3-- Improved roadside treatments that reduce the number of collisions with hazardous ;objects along the roadside.

• Goal 4-- Improved driver performance in run-off-the-road situations.

• Goal 1-- Optimum use of roadside safety features in relation to their selection, design, installation and maintenance.

• Goal 2-- Improved roadside safety hardware.

• Goal 3-- Improved vehicle compatibility and crashworthiness relative to roadside features.

• Goal 4-- Increased seat belt usage and effectiveness and enhanced occupant protection systems.

• Goal 5-- Improved emergency team responsiveness for highway accidents.

This paper will concentrate on providing a more detailed example to explain the process used to take a mission of the strategic plan and then narrow it down to specific actions. Obviously the main objective of the whole plan is reduce the number and severity of roadside accidents.

The strategic plan has the five missions. The first mission is focused on the need to create an awareness among the public about roadside safety, to help decision makers understand the problem, and to get the needed financial and political support to address the problem. Second, there is a need to develop the data and analysis procedures that are needed to fully understand the problem, monitor changing conditions, and come up with solutions. The third step is to keep the vehicles on the roadway by promoting smooth traffic flow, good maintenance, and proper driver behavior. If that can be done, then the problem of roadside accidents is solved. Practically speaking, it isn't always possible to keep the vehicle on the road, so, mission four is to keep the vehicle upright or keep it from hitting a harmful object, if it does go off the roadway. Finally, if rollover crashes cannot be avoided, then there is a need to minimize the injuries and the fatalities through effective design of the roadside and vehicle.

These missions, from an overall perspective, are logical, but they do not provide clear directives on actions that can be taken or provide a basis for measuring the effectiveness of efforts to address the roadside safety problem on each of these fronts. The structure used in formulating the strategic plan is based on missions, goals, objectives, and actions. The mission is a broad statement of an area to attack-like keep the vehicle on the roadway. Then within each mission there are specific sets of substrategies of how to approach that problem. These are reflected in goals which highlight the desired outcomes of the efforts. Below each goal, are more specific objectives. These are more detailed definitions of the approaches to achieve a particular goal. Finally, there are the things that specific agents or groups can do to address the problem-the activities or actions.

In the strategic plan, actions can be categorized into four different areas - research, operations, policy, and training. Research activities are obvious. These are activities oriented to getting the answers to questions about impacts or the need to get more basic information. The operations activities cover those things that can be done now; agencies just have to do it. It may take resources, but it can be done in the short term. Policy activities relate to changing legislation or policies within agencies to effect change more broadly. Finally, training activities are needed to train the engineers or other people in the DOTs and the public–the drivers–to actually behave more safely.

To demonstrate the detail in the strategic plan, I chose to concentrate on mission three– keep the vehicle on the roadway. An iterative process was used to define the goals, objectives, and actions under this as well as other missions.

There are three fundamental aspects that have to be considered in addressing the roadside safety problem— the roadway, the vehicle and the driver. These three aspects served to provide the basis for formulating the details of the strategic plan. The goals defined under mission three reflect these aspects:

• Goal 3.1 - Improved highway designs that reduce the probability of vehicles leaving the roadway.

• Goal 3.2 - Improved traffic operating environment that reduces the occurrences of events contributing to loss of vehicle control and roadside encroachment.

• Goal 3.3 - Sufficient maintenance of highways and vehicles to reduce the probability of loss of vehicle control.

• Goal 3.4 - Improve vehicle-based systems to help keep drivers on the road.

• Goal 3.5 - Improved driver performance and behavior.

The next step in the process involved defining objectives for each of these five goals. The objectives that were defined for the Mission 3 goals are:

• Goal 3.1 - Improved highway designs that reduce the probability of vehicles leaving the roadway.
  • Objective 3.1.1-- Develop the tools to allow highway designers to incorporate safety into the design process.

  • Objective 3.1.2-- Enhance design policies and guidelines to include safety considerations.

  • Objective 3.1.3-- Enhance the highway designer's understanding of the effects of highway design on roadside safety, in particular, the importance of consistency in highway design in aiding drivers to maintain vehicle control.

• Goal 3.2 - Improved traffic operating environment that reduces the occurrences of events contributing to loss of vehicle control and roadside encroachment.

  • Objective 3.2.1-- Manage operating conditions to reduce the probability of out-of-control vehicles.

  • Objective 3.2.2-- Improve signing, lighting, and delineation of the roadway to aid drivers' ability to stay on the road.

• Goal 3.3 - Sufficient maintenance of highways and vehicles to reduce the probability of loss of vehicle control.
  • Objective 3.3.1-- Maintain roadway elements adequately to minimize run-off-the-road accidents.

  • Objective 3.3.2-- Maintain vehicles adequately to minimize malfunctions which result in run-off-the-road accidents.

  • Objective 3.3.3-- Minimize the number and durations of work zones.

• Goal 3.4 - Improve vehicle-based systems to help keep drivers on the road.
  • Objective 3.4.1-- Develop reliable vehicular lateral guidance systems.

  • Objective 3.4.2.-- Develop vehicle systems to improve driver performance.

• Goal 3.5 - Improved driver performance and behavior.
  • Objective 3.5.1-- Change driver behaviors that contribute to run-off-the-road crashes.

  • Objective 3.5.2-- Support legislative and policy activities to improve driver behavior.

Under Goal 3.1 there are three objectives that focus on design approaches to reducing the probability of a vehicle actually leaving the highway. Developing better tools for the highway designers to analyze and incorporate safety represents one aspect of keeping vehicles on the road. Since such tools do not exist, the research at FHWA to develop the interactive Highway Safety Design Model (IHSDM) is a step in that direction. The next objective is to improve design policies and guidelines in order to keep the vehicle on the highway. Recent research has found that highway designers do not explicitly consider safety when following current design guidelines. A key reason for this is some rather significant gaps in understanding the relationship between design elements and their safety impacts. There are practices, such as safety audits or reviews, that attempt to assure that new road designs are safe, but more can be done to improve guidelines, tools, and policies.

The fourth goal under mission three is aimed at the vehicle-based systems to keep vehicles on the road. Two different objectives have been identified. The first objective is associated with the prospect of smarter vehicles that will have reliable vehicle lateral guidance systems. These systems would automatically keep vehicles on the road. The second objective is to give more information to the driver to help him or her keep the vehicle on the road. These might be improved signing, marking, or delineation treatments, or audible warnings that you are getting too close to the edge of the road or advance warnings of road narrowing or curves. Some of these things that are not available yet, but are expected to be in use over the next ten years.

The fifth goal is oriented to the driver. Three objectives have been defined for improving driver performance. The three objectives here address changing behaviors, policies, and driver education relative to roadside safety. Clearly, changing behaviors like the aggressive driver is not easy. Changing the driver’s behavior could be considered a training effort, but there is also research needs associated with it. Similarly, policy changes, legislation, and better enforcement or seatbelt laws can play a role in improving roadside safety.

Under the goal 3.3, the maintenance aspect is addressed. There are three objectives which indicate how maintenance can contribute to improving roadside safety. The objectives address maintaining the roadway (e.g., preventing shoulder edge drop-offs), maintaining the vehicle (e.g., replacing defective head lamps), and minimizing the number and duration of work zones which can be considered special cases of the roadside.

Then finally, in goal 3.2, there are two objectives. The first relates to the overall operating environment of the roadway, vehicle, and driver. Through effective management of the roadway operating conditions it is possible to reduce the number of vehicles going out of control and potentially leaving the roadway. Improvement in signing, lighting, and other aspects of delineation and guidance can be used for these purposes.

Through the efforts of the panel and many other professionals who participated in workshops or provided written comments on the strategic plan another level of detail evolved. For each objective a series of actions has been defined. These represent the specific elements of work or research that can contribute to meeting specific objectives and goals. It is important mention that the strategic plan has more than 350 action items. Full implementation would require millions of dollars to implement. This implies the need to do a cost effectiveness analysis, to see which ones of these actions should have the highest priority. This level of detail allows the entire plan to be stratified in several useful ways, and including:

• Action agendas by discipline

• Summary of research needs

• Summary of issue

There is not time to address all of the actions under mission three. It is possible to get a sense of this level of detail by looking at the actions under Goal 3.1 related to improving highway design procedures. The actions indicate a need to have a better understanding of the relationship between the effects of geometry and roadside safety. These are:

• Goal 3.1 - Improved highway designs that reduce the probability of vehicles leaving the roadway.
  • Objective 3.1.1 Develop the tools to allow highway designers to incorporate safety into the design process.
    • Action 3.1.1.1 Develop a better understanding of the effects of highway geometric design on roadside safety (e.g., sight distance, superelevation, curvature).

    • Action 3.1.1.2 Develop techniques to promote consistent designs that conform to driver expectancy.

    • Action 3.1.1.3 Develop improved hazard identification tools to identify potentially hazardous roadside designs and features on all roadways including local roads.

    • Action 3.1.1.4 Investigate the use of safety audits in the roadway design process.

    • Action 3.1.1.5 Develop uses of 3- and 4-dimensional visualization technologies to improve the design of highways.

  • Objective 3.1.2 Enhance design policies and guidelines to include safety considerations.
    • Action 3.1.2.1 Develop specific design guidelines for identifiable functional highway classes that consider the interrelationships among highway function, design, operations, economics and safety.

    • Action 3.1.2.2 Update design policies to recognize current vehicle fleet characteristics and operating speeds.

    • Action 3.1.2.3 Develop guidelines for making tradeoffs between safety and other project considerations (aesthetics, historic, etc.).

    • Action 3.1.2.4 Integrate new safety design tools and policies into existing roadway design software systems.

  • Objective 3.1.3 Enhance the highway designer’s understanding of the effects of highway design on roadside safety, in particular, the importance of consistency in highway design in aiding drivers to maintain vehicle control.
    • Action 3.1.3.1 Develop training materials for roadway designers, installers and maintainers that emphasize the relationships between roadway design and safety with emphasis on 3R/4R and similar projects.

    • Action 3.1.3.2 Develop training programs targeted for inexperienced engineers at the federal, state, and local levels and for consultants and design reviewers.

    • Action 3.1.3.3 Develop and maintain continuing education requirements and materials for experienced engineers.

    • Action 3.1.3.4 Encourage faculty to upgrade university programs to include consideration of roadway/roadside safety issues.

This would allow a more consistent design process. Research actions are needed to improve hazards identification, and to integrate safety audits, and other procedures that could be used to enhance the design process and assure that designs are effectively implemented. The actions also suggest that 3- or 4-dimensional visualization technologies will provide a driver's perception of the roadway to help designers assess the effectiveness of designs, traffic controls, and other features that affect roadside safety. Other actions address the need to convey new procedures to the actual people designing the systems. There must be simultaneous actions to get the policies changed, so that all this new information and procedures get used by the highway agencies; that sometimes can be a very difficult and slow process.

The present version of the plan has been through seven iterations starting from the very broad missions to the specific actions just described. It has been a tedious process with input from many professionals that led to this plan comprised of 5 missions, 25 goals, 121 objectives, and 358 action items. There has been constant editing, consolidation, and reorganization, but the panel is satisfied that the strategic plan comprehensive and useful. The remaining challenges will be to implement the plan, set priorities for individual elements, and continually update the plan to reflect changing conditions.

IMPLEMENTATION AND RESEARCH NEEDS
Kenneth S. Opiela
TRB, Senior Program Officer

Introduction

In an effort to address the roadside safety problem, the National Cooperative Highway Research Program assembled a distinguished group of experts and charged them with the task of identifying ways to improve roadside safety. In their deliberations, they formulated a vision, missions, goals, objectives, and actions for improving roadside safety as has been described by the previous speakers. It is important to note that considerable effort has gone into the development of the strategic plan. The project panel, several consultants, and many others, including those from very different backgrounds and diverse perspectives on safety, have contributed to content and organization of the strategic plan through numerous meetings and workshops. It has been estimated that over 2.5 person years of time was contributed in meetings and consultant efforts for the plan. The strategic plan has reached a level of considerable detail as a result of the incremental build, review, and revise process that was utilized.

The mere existence of a strategic plan does not assure that roadside safety will be improved. Further, neither the NCHRP nor its expert panel have the authority to implement such a plan. Recognizing these facts, the panel has attempted to develop recommendations for future directions in roadside safety using the strategic plan. It was recognized that the plan can be thought of as a comprehensive list of strategies or actions that can contribute to improved roadside safety. Some of these strategies are regularly being done, while others may be possible but not getting adequate attention. There are, of course, those things that cannot be done because relationships are not understood, agencies are not communicating, data is lacking, or systems have not been developed. Based upon these distinctions, the panel has been pursuing the development of implementation recommendations and lists of research needs.

The five fundamental missions and the associated goals of the strategic plan are presented in Table 1. The subdivision of the missions and goals into objectives, actions, and needs allows the roles of various agents to be highlighted, action agendas formulated, and needs for research identified. The following paragraphs highlight some of the actions that can be effective, the implementation efforts that are necessary, and areas where research is needed.

The efforts to identify actions in the strategic plan provided the basis for defining specific activities that could be undertaken to improve roadside safety. Summaries of these actions by discipline provide an action agenda that can reinforce continuance of current activities or to highlight other things that can be done. Lists of action items are being developed for the planning, design, operations, maintenance, and administrative functions of DOTs. In addition, the strategic plan is providing a basis for identifying roles for law enforcement, driver training, construction, vehicle design, and other related disciplines. These action plans provide a basis for systematic review of an agencies efforts to address roadside safety problems.

The details provided in the strategic plan also provide the basis for further analyses. First, decomposition of the strategic plan also reveals areas where synergy can be achieved through the coordination of actions. It is also possible to isolate all actions that can be directed to specific problems for program development purposes. For example, crashes with trees are known to be a serious problem. By tagging all elements of a plan that have direct relationship to an issue, it is possible to recognize the necessary aspects of a program to address that issue. The strategic plan also permits interaction analysis to provide the foundations for the establishment of coalitions, linking of activities, and finding places where multiple purposes can be achieved. It can facilitate the consideration of the trade-offs between actions or agents to determine the most effective.

The NCHRP effort identified many actions for addressing roadside safety problems. Some examples of important actions are:

• Install shoulder rumble strips to alert drivers-- It is becoming apparent rumble strips located on the highway shoulder are effective in alerting drowsy or inattentive drivers when their vehicle is drifting off the roadway. These have been effectively installed on some interstate highways, but their potential use extends to rural two lane roads as well. Rumble strips represent an example of a relatively low-cost remedy for certain roadside safety problems.

• Strategically remove or shield trees or utility poles close to the roadway-- Every day people are killed because of crashes into trees and utility poles located along the roadway. Pole placement policies need to be re-examined and programs need to be developed to reduce the number of poles on the roadside. Trees are even a more common roadside hazard than poles. Strategies for removing trees in particularly hazardous locations on the roadside (e.g., on the outside of tight curves) are needed to enhance the safety of the traveling public while maintaining the aesthetics of our highways.

• Use public service announcements & citizen initiatives to increase awareness-- Impressive reductions in DUI have resulted from citizen initiatives such as MADD and SADD. Similar actions are needed to attack the roadside safety problem such as increase awareness of roadside safety, encourage greater use of seat belts, or discourage excessive speed on curves.

• Improve safety management systems-- Transportation agencies need better data and tools to manage highway safety. These systems can improve the practices for identifying hazardous locations, link highway features and accident data, assistance in making optimal decisions related to limited resources, and allow monitoring of changing traffic conditions that might increase roadside crashes. With almost 4 million miles of highway in the US, it is a major challenges to monitor safety. While most agencies maintain accident records and highway inventories, there is a need to upgrade, expand, and link this information to increase basic capabilities for safety management on a continuing basis.

• Implement proactive highway maintenance programs- Adequate maintenance will assure that the surface provides the skid resistance necessary to negotiate curves, minimize erratic maneuvers to avoid potholes, and eliminate shoulder drop-offs that can lead to loss of control and departure from the roadway. Part of maintenance attention should be the maintenance of adequate sight lines for the driver.

• Improve driver education programs-- Drivers need to learn about the hazards of the roadside and rationale of traffic controls to help them to avoid situations where they could leave the roadway. It is even becoming possible to use simulators to train drivers in maneuvering through unusual situations, such as leaving the roadway.

• Solicit support in identifying hazardous roadside locations - The public, police, main-tenance crews, and others are often aware of roadside situations that are hazardous. Agencies should find means to solicit input from these groups to develop a comprehensive inventory of potential roadside hazards to supplement information derived from crash reports. This would allow an agency to take action before it becomes a problem.

• Provide guidance on the use of effective roadside safety hardware - Persons responsible for highway design and safety management need guidance in the selection and application of roadside safety hardware to assure that it is used where needed and installed so that it will function effectively. The Roadside Design Guide and other such documents need to be kept up-to-date, field installations done properly, and in-service performance monitored to have a positive effect on safety.

• Promote proper installation and maintenance of roadside safety features - Safety features are effective only if they are used appropriately and installed properly. "Fool-proof" methods of installation are needed to assure that roadside hardware functions as intended. Roadside maintenance management systems should be implemented to assure prompt repair and appropriate periodic maintenance

• Increase speed enforcement at locations with known roadside safety problems - Plan police enforcement activities at or near locations where roadside crashes are known to occur. Use automated enforcement methods when they are available to increase enforcement coverage.

• Promote development of innovative technologies to keep vehicles on the road - Utilize ITS technologies to develop systems that will help drivers stay on the road or avoid hazardous objects or automatically prevent collisions.

• Improve the proficiency of persons responsible for roadside safety-- Support safety professionals with a legacy of knowledge, pervasive, interactive networks for information exchange, and national databases. Upgrade training courses and use different means to facilitate their use.

• Improve vehicle design to increase compatibility with roadside hardware-- Upgrade vehicle design standards to assure compatibility to address multiple needs (e.g., establish minimums and maximums for bumper height to improve safety in both vehicle-to-hardware and vehicle-to-vehicle crashes).

• Gather and maintain improved data for safety analysis-- Implement systems to provide more accurate accident locations. Develop and maintain inventories of roadside safety features and maintain records of roadside hardware maintenance. Develop procedures for monitoring reports of roadside accidents.

• Improve hardware design-- Develop improved methods for the design and testing of safety hardware including simulation methods. Exploit the properties of new materials in roadside safety hardware and review opportunities for new hardware. Assess the functionality of hardware in field applications.

Obviously, some of these actions can be undertaken immediately, at little cost and others will take time and larger amounts of money to implement. There is a strong consensus, however, that these and other actions can lead to major reductions in the deaths and injuries that are occurring on the roadside. These actions and others are described more fully in the report being prepared by the NCHRP.

During the efforts, it was recognized that highway safety encompasses a very broad range of organizations including:

• State and Federal Departments of Transportation,

• Local law enforcement agencies,

• Emergency services providers,

• Citizen action groups,

• Automobile manufacturers, and the

• Insurance industry.

Each of these groups has its own specific areas of expertise and concern which sometimes complement each other and other times work against each other. The primary purpose of a strategic plan for roadside safety is to form a framework to unite all these different organizations in coordinated action for improving the roadside.

Implementation of the strategic plan will require addressing the resource, procedural (process), and administrative issues listed below through coordinated efforts with the various organizations:

• Increased funding is needed to address the many known problems.

• Improved information systems and means to capture data are needed to effectively monitor roadside safety and select appropriate solutions for locations with problems.
  • Old computer systems and legacy data are difficult to move to newer computer platforms.

  • While many focus on the need for better accident data, highway agencies must also significantly upgrade their inventory and traffic data resources.

  • New technologies need to be employed to facilitate the capture of data about crashes and to make that information available to those responsible to maintaining safety.

  • There are many gaps in knowledge that make it impossible to assess the effectiveness of roadside treatments.

  • Facilitate the availability and exchange of information through the use of advanced technologies or the establishment of information clearinghouses.

• A champion or coalition is needed to focus on this issue and promote continued attention.
  • A strong champion would be desirable to take the lead and coordinate efforts among the disciplines. A coalition approach could also be used if a consensus can be reached on specific roles.

  • Communication between partners and stakeholders is vital to making the plan work.

  • Coordination of activities is critical to assuring that the problem is addressed.

  • Periodic review of programs and processes is needed (safety audits).

  • Enhanced safety management systems are required for comprehensive network-level evaluation and monitoring of safety. These systems should be based upon timely data on accidents and information on roadway features.

  • Safety decisions should be based upon sound cost-benefit analyses. Convenient decision support tools should be developed and utilized. Analytical tools need to be enhanced and extended to allow routine application on a system-wide basis.

• Sound policies supporting safety efforts are needed to assure effectiveness.

• Permit flexibility in the use of safety funds.

• Information needs to be provided to decision makers to assure that they understand safety needs and to increase their awareness of problems and successes.

• Mechanisms are needed to initiate and support coordinated efforts among the various agencies that have a role in highway safety.

• Public liability issues need to be resolved on a broader level to avoid significant losses to public agencies.

The comprehensive review of the strategic plan to improve roadside safety led to the identification of areas where research was needed. Important research needs by mission are described below.

• Develop communications and training techniques to improve the awareness and knowledge of roadside safety hardware installation and maintenance personnel, local law enforcement personnel responsible for accident reports, and key decision makers.

• Develop mechanisms to create a feedback loop between the research community and federal, state, local and private partners to enhance the exchange of information concerning roadside safety research results and implementation experience.

• Investigate ways to communicate research results and training techniques effectively to local and state practitioners, targeting issues and approaches suitable for local roads and county and city engineers.

• Develop improved methods of collecting real-world data necessary to enhance the design clear zone and side slopes, and to better design, install, and monitor roadside hardware.

• Research aimed at developing increased knowledge of: a.) the effect of modifications to the clear zone and side slopes on crash rates and severity; b.) the in-field crash performance of roadside hardware components, including the effects of installation and maintenance practices; and c.) general measures of the severity of roadside hardware impacts as affected by vehicle type, speed, and angles of impact.

• Research aimed at developing tools for state/local roadside improvement programs, such as data-driven procedures for roadside safety decisions as part of the Safety Management System.

• Develop techniques to encourage consistent designs that conform to driver expectancy.

• Investigate the use of safety audits in the roadway design process.

• Research the effectiveness of shoulder rumble strips, delineation of roadside hazards and fixed objects (poles, trees, etc.), and signing and lighting systems for all conditions, including severe weather.

• Develop and implement guidelines for low-cost maintenance improvements on low volume roads to reduce run-off-the-road crashes.

• Develop in-vehicle systems to provide safety information to the driver and/or assist in avoiding collisions with roadside objects.

• Develop technologies to enhance driver visibility during nighttime and adverse conditions (head lamp design, ultraviolet headlamps, etc.)

• Research driver monitoring systems and define driver behaviors and characteristics that contribute to loss of vehicle control.

• Develop and deploy speed enforcement strategies which promote safer driving behavior.

• Improve understanding of interaction between vehicles and roadside geometrics to develop roadside terrain model and update guidelines for slopes, drainage, edge drop-offs, curbs/gutters, and other features.

• Develop improved benefit-cost analysis methodologies for the selection of roadside features.

• Evaluate performance of roadside safety features of side slopes and develop guidelines to minimize impacts. Revise warranting criteria to reflect new knowledge about roadside crashes. (e.g., median barrier warrants).

• Develop simulation programs based upon driver response to running off the road situations to generate information needed in driver training programs and evaluate the effectiveness of such training.

• Develop new methods for evaluating the risk of occupant harm in roadside hardware crashes. Assess the assumption of restrained occupants in full-scale crash tests.

• Improve crash testing procedures based on assessing many factors include the field relevance of current full-scale test impact conditions, site geometry for crash tests of guardrails, crash cushions and guardrail terminals, and the performance of a variety of vehicle types in roadside hardware crash tests. Examine possible linkages between vehicle compliance tests (FMVSS tests) and roadside safety hardware testing.

• Develop improved hardware or new hardware including generic guardrail terminals, guardrail terminals that perform acceptably in side impact collisions, and energy absorbing devices for trees and rigid poles. Use new and innovative materials in roadside hardware.

• Improve the maintenance of roadside safety hardware to assure that it will perform effectively when needed. Identify owner-agencies with innovative and effective methods of ensuring that roadside hardware is properly installed and maintained. Utilize maintenance equipment effectively to minimize the time and cost for safety hardware repair or replacement.

Roadside safety researchers have played a significant role in reducing fatalities and injuries in roadside collisions during the past several decades. In the past many problems could be solved with the intuition and judgment developed by researchers during 20 years of crash testing and observing accidents in the field. Many of the problems that remain are the difficult ones that have defied solution for many years. Maintaining the current level of safety in the face of increased travel demand, lighter vehicles, and strained public funding resources will require a bold strategic plan that coordinates the efforts of all agencies involved and maximizes the effectiveness of future research and development efforts. Roadside safety is a huge problem that can be significantly reduced.

The efforts under the NCHRP Project are believed to have laid the groundwork for further effort and importantly the development of a coordinated approach to improving roadside safety. More effort is needed, particularly to assure that all perspectives are considered and the most cost-effective options are selected. With the resources available for this project it is hoped that a more detailed strategic plan can be developed and momentum promoted towards the implementation of the strategic plan. The next step in the process will involve a series of meetings in April and May of this year to add detail to the strategic plan, particularly relative to defining goals, objectives, actions, and research needs. It is envisioned that three working meeting will be held with 15 to 30 persons representing the various interested organizations. These meetings will focus on particular missions develop the appropriate additional elements under each. A national meeting is planned for late summer or fall during which the entire strategic plan will be presented and reviewed. Working sessions at this meeting will assess the completeness and feasibility of the strategic plan and attempts will be made to prioritize the research needs, build consensus, and promote a champion for the strategic plan for improving roadside safety.

A great deal has been accomplished in improving the effectiveness of roadside safety hardware during the past several decades. The always-changing vehicle fleet and highway environment do not allow the roadside safety community the luxury of complacency. There are significant challenges ahead in improving roadside safety. These challenges can only be met by openly discussing difficult issues as they emerge and focusing the efforts all those with an interest in roadside safety on coordinated action.

Table 1 - Missions and Goals of the Strategic Plan

At first, a lot of people ask why an update is needed since NCHRP Report 350 just came out. It is important to note that NCHRP 350 was published in 1992. It was adopted (by reference) as the FHWA standard in 1993, but since the compliance date for safety hardware was set for August 1998, the report has not gotten much attention -- until lately. Further, the updating process for NCHRP Report 230, the predecessor document, took six years from the time they first started talking about it. It takes quite a long time to systematically review a technical document and prepare an update, but it takes longer and longer as the procedures become more comprehensive and complicated. The evolution of the process clearly indicates growing complexity. The first procedures for crash testing (Highway Research Circular 482) took about six months to develop and prepare and it was one page in length. As hardware safety testing evolved, it took two to six years to update every time. So, it is the right time to begin looking into updating NCHRP Report 350. One of the things that needs to be emphasized is that the current research under Project 22-14 is not intended to produce the updated document. It is intended to look at specific needs for an update and to identify the research that may be needed to provide a sound basis for the update. Frankly, the funds available do not provide enough money to update Report 350. The objectives of this research defined by the panel are to 1) evaluate the relevance of procedures for the safety performance evaluation of highway features, 2) assess the needs for updates to the NCHRP Report 350 document and recommend strategies for implementing them. It is assumed that this project may identify specific needs for immediate updates to NCHRP Report 350. Project funds have been reserved to provide resources for the production of an interim update, if necessary. It is also assumed that a full update will eventually be needed and additional funding will be necessary at that time.

When the panel first met, the members started with a long list of concerns about NCHRP Report 350. The following is not an all inclusive list of those concerns and it is not in any particular order. It does, however, indicate some of the things that the panel will think about for the update. These include:

• Test Vehicle-- There has been a lot of concern about using the pick-up truck. Is it the right test vehicle?

• Vehicle Age-- Are we using the right age of vehicle? When you consider that a test can be run with a vehicle up to five years old and then it takes about a year to get the certification after the tests are run, you are certifying hardware for fleet characteristics that are six-years out-of-date. Is the safety hardware still viable given that fleet characteristics have continued to change?

• Testing of Vans-- Should vans be tested? What is their percentage of the fleet? Changing fleet characteristics are a concern recognized by the panel.

• Test/Real Conditions-- Concerns have been expressed that test conditions do not reflect the real conditions in which roadside hardware is installed. Typically, crash tests are run on flat, level ground. Often hardware is installed on a slope because drainage concerns must also be addressed in highway design. Are these the right test conditions? Should tests be run for one or more basic ditch configurations? There are millions of ditch configurations, so it adds the dilemma of having to pick an appropriate one. There are other factors also associated with crashes involving vehicles trying to return to the roadway on a slope, but there is limited current information available.

• Acceptance Criteria-- Should partial rollovers be allowed? It is a concern that a test is considered a complete failure when it passes in every other aspect of the test but because a gust of wind flips a pick-up truck on its side it does not meet the pass criteria. There is a serious difference between a gentle roll and the multiple flips and rolls sometimes seen in crash films. Is that a sufficient reason to fail a particular piece of hardware?

• Test Matrix-- Do we really need all of the tests? The industry has noted that they are doing at least twice as many tests under NCHRP Report 350 and consequently it is driving the cost of hardware up. To give you an example, in the state of Illinois about 1,200 end treatments are installed a year. They used to cost $800 to install, but they now cost $3,000 to $4,000. That represents a significantly higher cost to the agency in times of limited budgets. With the higher cost of a full-scale crash tests, do we need all of them? Somebody has to pay for these tests, and in the case of the DOTs, it is the taxpayers.

• Use of Simulation-- Could more be done using a simulation testing? Is it possible to substitute some of the full-scale tests with simulation testing?

• Occupant Restraint-- The President said recently that he wants to get more people using occupant restraints. Do we need to worry about the guy who says, "I don't care to be in a seat belt" when testing roadside hardware? Maybe crash testing should be done assuming a restrained occupant. Maybe the worst case condition isn't needed all the time.

• Test Replication and Standardization-- Some researchers involved with crash testing will acknowledge that test results can be influenced by test set-up. Soil compaction, moisture content, material and component choices, and vehicle selection can be used to influence a test to pass or fail. Is there a way to standardize the tests more?

• International Harmonization-- There needs to be some standardization between testing in the U.S. and other countries, to accept products or understand their performance range.

Since most of the participants here don't represent state DOTs, it is appropriate that state DOT concerns be expressed. The DOTs are concerned about the costs of roadside safety hardware. About two years ago, the FHWA told the DOTs not to install breakaway cable terminals anymore. In Illinois, it was decided that products tested to NCHRP Report 350 would be used instead. This, of course, generated flak throughout the state DOT because of the increased costs. It is a big issue in a state like Illinois, where the DOT has 17,000 miles of roadway under direct supervision. A backlog of about 2,800 miles of roadway that needed repairs existed, but couldn't be funded. That backlog is anticipated to grow to around 4,000 miles over the next five years to some extent as a result of higher costs to meet safety requirements. It is not easy to convince the director that increased spending on guardrail end treatments is needed. The director doesn't get a lot of phone calls complaining about a outdated guardrail end treatments, but he gets a lot of phone calls about potholes. Looking at the Illinois in-service inventory, one notes roadside hardware treatments that have been in use for years. The DOT has a stockpile of components, maintenance people know how to put them in place, and designers know where they are effective. The need to meet a higher requirement all of a sudden renders these things outdated. Is there a need to include in the testing procedures a caveat for devices that have acceptable, long-term in-service performance. Is there a need to provide for less stringent in-service testing? In some cases, there is hardware in Illinois that didn't pass 230, and doesn't stand a ghost of a chance of passing the Report 350 criteria. Yet, the accident data indicates that it is working. There is also some hardware that passed 230, but it didn't work which in-service performance evaluations identified. There is a need to include acceptance criteria for hardware with effective in-service performance without crash testing. The states need a way of accepting products that have been used for years and that have been proved with experience. It seems rational to assume that no state would continue to put in hardware that was known to be killing people.

Reviewing the roadside hardware standards from the various states -- I used to work in the standards department and I have standards books from about every state in the United States -- one will note that there are more than 200 guardrail designs. Agencies forced to change their basic designs will face other difficulties. For example, Illinois is a steel post, steel block-out state. There are probably millions of posts and block outs installed. The wood industry has approached the department a couple of times about coming in. The DOT has said that they will accept wood post and wood block outs. Selling the alternative to contractors is another matter. The contractors argue that they know how to install steel posts and steel block outs and are equipped to install them. Changes will be costly to them. Other considerations are whether the test labs can complete testing to meet the 1998 date. Higher costs can be expected in other areas as well. For example, temporary concrete barriers are used extensively in highway construction projects across the country. It hasn't passed Report 350 yet. What are we going to do with all of the non-standard concrete barriers which because of their durability would be expected to get many more years of use in construction? These are major issues to the state DOTs.

The panel determined that an important aspect of this project is the relevance issue. The term often gets mentioned, but there is no universally accepted definition. The panel probably spent more time discussing relevance and trying to define relevance than on any other part of the work statement. The definitions and thinking changed frequently during the meeting. But the biggest concern of the panel that remained throughout was, when a vehicle leaves a roadway and hits a barrier, is it effective. All the crash tests and all the data isn't any good if lives are not saved and injuries reduced. After that comes the question of whether the prescribed crash tests relate to the types of hits that are occurring and at the levels of severity assumed. Good data is limited, but essential to providing these insights.

We need to address the relevance issue based on available data at this time, because there is a lot of data that is needed. Part of what we are hoping to get out this effort is how to get at that data. What research is needed.

Under Project 22-14, the contractor is undertaking a work plan that will involve the following efforts:

• Conduct a literature review and also a survey. The survey will aim to contact about persons representing selected DOTs, industry, safety advocacy groups, and others to assure that all updating needs and issues are identified.

• Develop prototype methodology and establish criteria for determining relevance. The data needs for the methodology will be defined.

• Conduct a preliminary assessment of the specific needs for updating to determine the feasibility and impacts.

• Recommend updating efforts and strategies. Meet with the panel to decide which recommendations will be pursued.

• Fully develop relevance assessment methodology.

• Conduct investigations, approved by the panel, to establish the background for recommendations for updates to NCHRP Report 350. Assess the need for an interim update.

• Request outside review of the recommendation for updating NCHRP Report 350 to test the feasibility of the recommendations.

There is a little bit of money in our budget for a possible interim update if the panel identifies something that really needs to get out there now. This should help the industry people to relax a bit relative to the potential of changing requirements.

WORK PLAN UPDATING FOR PROCEDURES FOR SAFETY PERFORMANCE EVALUATION
King K. Mak
Texas Transportation Institute

The NCHRP Project 22-14, "Improvement of the Procedures for the Safety Performance Evaluation of Roadside Safety Features," research effort is divided into two phases and 10 tasks. The project started in March of this year and is proceeding as planned through the first phase of the work. The task elements are described below.

Task 1 involves a literature review and contacts with safety professionals. This meeting was planned as part of the effort to solicit inputs on updating needs and the relevance of the testing guidelines from more than the academic or research side. It is expected that the participants of this meeting will have the opportunity to provide insights and specific information, and openly propose issues that should be considered in the updating process. These inputs will be combined with the findings of the literature and the feedback from other contacts with professionals at the Federal Highway Administration, the state DOTs, universities, and industry to provide a comprehensive list of potential updating topics. These will be incorporated into an interim report to the panel from which they will pick specific items to be addressed in the second phase of the project that will lead to recommendations for future updates.

Task 2 attempts to develop a prototype methodology to assess the relevance of testing procedures. There is not enough money to fully address this issue, so it has been proposed that first a definition of relevance be established and then viable measures of it be determined. The research team has tried to deal with the definitions of relevance and how it can be measured to stimulate discussions during this meeting. It is believed that relevance measures are critical to ascertaining the need for a test outlined in the test matrix as well as to provide a means to identify emerging safety problems with roadside features. After the meeting, the research team will attempt to devise a prototype methodology that can be used in the future to address this issue.

Task 3 will focus on the assessment of the updating needs. A number of concerns have been identified, including:

• Test Vehicles-- Current test requirements are for tests with an 820 kilogram passenger vehicle at the low end and a 2000 kilogram pick-up truck at the high end. There has been considerable changes in small cars over recent years. The curb weights of the most common models has crept upward making it increasingly difficult to meet the test requirements. What will the nature of these small cars be six to ten years from now? Another issue is the adequacy of the 2000 kilogram pick-up truck as a surrogate for the entire light truck class of vehicles. There has not been testing to confirm that the different vehicles within the light truck class perform in a similar manner. Another concern that has been raised is that testing uses vehicles at the two ends of the spectrum (i.e., a very small passenger car and a pick-up truck). Should a medium-sized vehicle also be required? There are also questions about how representative these vehicles are relative to real-world accidents. Current testing may be using the wrong vehicles.

• Impact conditions-- Current tests require that specific impact angles and speeds be used. Are the impact conditions of 100 kilometers per hour, and 25 degrees, for example, a good set of test conditions? Should the impact angle be reduced from 25 degrees to maybe 20 degrees, or should there be an increase in speed from 100 kilometers per hour to 110 kilometers per hour or even 120 kilometers per hour to reflect the higher speed limits on the highways?

• Test Matrices-- A set of tests or a matrix is specified to establish the crashworthiness of a devices or structures. Are there too many or too few tests required? Are tests based on the right conditions to reflect the nature of impacts that occur in the real world?

• Evaluation Criteria-- NCHRP Report 350 defines specific evaluation criteria for determining whether a particular tests is considered a success or a failure. The criteria are both quantitative and qualitative, and many questions have been raised about the criteria. For example, the criteria specifies that there be no rollover if a test is to be passed. There have been numerous tests ran where relatively gentle, 90 degree rolls occurred. It has been suggested that these rolls should be allowed for some types of tests. There are questions about the efficacy of the flail space model and the ride-down deceleration criteria. Are they effective measures, or should emerging technology be used to implement evaluation criteria that are a little bit more sophisticated? Maybe there are other criteria that are not well defined that can be improved. For example, for pick-up trucks, there is often intrusion in the passenger area resulting from the buckling of the floor pan. Currently, there no quantitative measures of how much intrusion is too much, but it might be possible to incorporate an improved evaluation criteria for intrusion.

• Underlying Assumptions-- There were many assumptions made in formulating the guidelines in NCHRP Report 350. These underlying assumptions need to be reconsidered in the light of current and anticipated conditions. For example, the issue of restrained versus unrestrained occupants is raised often. Right now, tests assume unrestrained occupants. With the increasing seat belt usage, it may be appropriate to consider restrained occupants and alter the evaluation criteria accordingly.

• Instrumentation-- Is the instrumentation required in test vehicles and hardware sufficient? Is there a need for more or different instrumentation to allow more reliable measurements or other measures of performance?

• International Harmonization-- NCHRP Report 350 was prepared considering the need and value of harmonization between the United States, the EU as well as other countries including Australia or even some of the Asian countries that have started to move in this area. The methods and criteria used by CEN and other countries may be useful for the updating effort.

• Test Information Requirements-- There has been a lot of discussion about documenting how prototype devices are built and installed for testing. Current requirements are pretty loose. Are more stringent requirements needed?

• Soil Type/Condition-- Many believe that soil/type condition has a big effect on the outcomes of crash testing. It would be useful, but possibly costly, to standardize soil type and condition for testing. There is a need to look at the options for improving controls for this aspect.

• Test Documentation and Reporting-- There are often complaints that there is not enough information provided in the test reports. What documentation should be added or altered? Should failed tests be documented?

• Alternatives to Full-Scale Crash Testing-- Full-scale crash testing is costly and it has been proposed that other analytical and experimental techniques be used to reduce the costs. Has computer simulation and modeling been developed to the point that it can be used to replace some crash testing? Should there be any requirements for these techniques to be incorporated into the update?

The research team will look at these topics on a preliminary basis try to provide background materials, assess the impacts and updating needs, and make recommendations in an interim report.

Task 4 will involve a meeting with the project panel. The interim report will serve as the basis for detailed discussions at a panel meeting to determine the importance of the issues and to set priorities for changes to be incorporated in an update to NCHRP Report 350.

Task 5 will develop more detailed procedures for determining relevance. The results of the pilot studies and discussions of the interim meeting will allow a more in-depth review of means to measure relevance.

Task 6 will begin the process of updating by focusing on specific issues preparing "white papers" on these subjects. These white papers will present relevant background material, offer draft versions of the proposed changes to Report 350, provide commentary on the rationale for the changes, and assess the expected impacts on topics ranging from costs of testing to highway safety.

Task 7 will try to come up with some strategies for updating the NCHRP Report 350. These strategies will consider various scenarios of timing and scope for the updates. In essence, the panel will look at the items that need to be updated and determine which should be addressed right away. The project has funds for an interim update, if that is determined to be appropriate. A time table for undertaking further analyses and research, including crash testing, will also be formulated to assure a sound basis for the full update of the procedures.

In addition, updating issues not covered under the current contract will be prioritized. The research team will outline those topics that should be addressed prior to the full effort to update NCHRP Report 350. These efforts may involve further research efforts, crash tests, and/or integration and validation of findings from other research.

Task 8 will focus on the preparation of a second interim report summarizing the findings of Tasks 5 through 7. This report will provide the panel another opportunity to consider updating issues in more detail.

Task 9 will involve an outside review to solicit feedback on the proposed updates to NCHRP Report 350. It is not clear how this review will be handled. It might be possible to use the next summer meeting of A2A04, or it may be distributed for review and comment via the mail. The intent will be to provide parties another chance to offer arguments about the nature and need for proposed updates to the procedures for safety performance evaluation.

Task 10 will complete the efforts under the current project and document these efforts and the findings in a final report.

The planned schedule for Phase one, which is tasks 1 through 4, should be completed by September 1997. Basically, position papers have already been drafted. Inputs from this meeting will be incorporated into the position papers, and these will be put together into a report by September. The project panel will be provided the interim prior to an interim panel meeting before the end of 1997. In the interim panel meeting, the project panel will address differences of opinion, confirm or alter the topics that are considered important for an update by the project staff, and establish consensus on the priorities for addressing the issues. It is expected that efforts to develop the white papers will be underway in early 1998. The critical (second) review will begin sometime in mid- to late 1998 with the expectation that the project will end in May of 1999. Since the need for an interim update will not be determined until the interim meeting, there is no schedule for when that might occur. It is, however, expected that a full update should be planned to occur by 2002, which would be ten years after the publication of NCHRP Report 350.

In order to define and measure the relevancy of crash-testing procedures, we basically need to know three things about those procedures:

• Are the procedures reliable?

• Are they valid?

• What are the consequences that flow from the procedures?

Let me begin by distinguishing between reliability and validity. In the shot pattern shown in figure 1, we see that the shooter is off target. More importantly, we see that the shot pattern is scattered all over the target.

FIGURE 1: UNRELIABLE SHOT PATTERN

There is no consistency, no reliability to the shot pattern.

In the second figure, the shot pattern is much tighter than in the first. The shooter is still off target, but at least the shooter is consistently off target. The shot pattern is reliably concentrated in a small area.

FIGURE 2: RELIABLE SHOT PATTERN

In the third figure, the shot pattern is again consistent, but in this figure the pattern is also concentrated in and around the bull’s eye. This pattern is valid. Validity in this case means that the shooter is actually hitting that which he or she intends to hit, the bull’s eye.

From just these three figures, it should be clear that "validity implies reliability," but reliability does not necessarily imply validity. If a given crash test is repeated many times, and if the results of each test are consistent, then the test is reliable, though not necessarily valid.

Reliability, as the term might be applied to crash testing, can actually be divided into two categories: intra-facility reliability and inter-facility reliability.

• Intra-facility reliability means that repeated tests of the same appurtenances under the same test conditions at the same test facility yield the same (or similar), outcomes.

• Inter-facility reliability means that the same tests run on the same appurtenances under the same test conditions at different test facilities, yield the same (or similar) outcomes.

With sufficient funding it would, obviously, be possible to assess directly the intra-facility reliability of a particular crash test procedure by replicating a given test several, say, five or ten times. To the extent that all of the tests yield the same or similar results, the reliability of the test procedure is supported. [Unfortunately, when roadside appurtenances are tested for compliance with a set of guidelines, very few (if any) replications of successful crash tests are ever run. Once an appurtenance passes a test, that test is not typically repeated.]

In similar manner, the inter-facility reliability of a given crash test procedure could be supported through "round robin" trials in which several crash test facilities all conducted the same crash test. To the extent that all participating facilities produced the same (or similar) results and arrived at the same conclusions, the inter-facility reliability of the testing procedure would be supported.

As a practical matter, I doubt that funds will be forthcoming to test the reliability of current crash testing procedures. In the absence of funds to directly assess the intra- and inter-facility reliability of crash testing procedures, some estimates of the reliability of existing crash testing procedures might nonetheless be undertaken indirectly.

To indirectly assess the reliability of crash test procedures, we would start by identifying those factors that are fairly critical to the reliability of the tests (i.e., those factors that impact the outcomes of individual tests). Some of the factors impacting crash test reliability are shown in table 1.

If we are going to have reliable crash tests, instrumentation is a major area of concern. Are the outcomes of individual crash tests impacted by differences in transducers, telemetry systems, on-board recording systems, filters, etc. And if these instruments are out of calibration, how is the reliability of the tests threatened?

When we reduce crash-test data, at least across organizations, different protocols may be employed. These different protocols may add to the variance in our problem, to the variability in the perceived outcomes of crash tests.

Documentation and reporting do not directly impact the reliability of crash tests, but if we are ever to understand the differences in crash-test outcomes, we have to have reports that are well documented and fully explained. NCHRP 350 does a pretty good job of outlining the documentation and reporting requirements for crash tests, but I think this area is deserving of more attention.

What about the test vehicles themselves? Is there any variance there? Some vehicle parameters are certainly specified in NCHRP 350: mass, wheel base, and so forth. But, we do not have a requirement for bumper height, suspension characteristics, optional equipment, inertial properties, or some measure of the frontal profile of the vehicle. These are characteristics that are left at variance—characteristics that can affect the outcomes of these tests.

Under the provisions of NCHRP 350, engineers are left with options for some tests. One example: the testing of break away support structures at 35 and 100 kilometers per hour (tests 360 and 361). These tests can be run so that the striking vehicle impacts the test structure "dead center" or midway between the centerline of the striking vehicle and the left- or right- outside edge of the vehicle. This option gives the investigator some latitude while again allowing more variance into our problem.

What about subjective evaluation? NCHRP 350 defines some outcome measures fairly well. Occupant impact velocity and ride down acceleration, for example, are both well defined that are mathematically quantified. But what about occupant compartment penetration or deformation, debris, and vehicle stability? It seems to me that we need some good operational definitions for these more subjective measures of test outcome. We do not define these terms nearly as explicitly as we might.

Test installations: NCHRP 350 gives some tolerances for conducting tests (e.g., speed and angle) but relatively little on the test installation itself. In some cases these discrepancies from the nominal (intended) height, width, strength, etc. may impact the outcome of the test.

Finally, there are variables (e.g., soil compaction, moisture) that are unaccounted for in NCHRP 350. When these variables are not accounted for, as in the previous examples, variance (unreliability) can be introduced into our problem. Two identical guardrails run under the same test conditions (except for the characteristics of the soil in which the guardrail posts are embedded) may yield very different test results.

In summary, if we want to run more reliable crash tests in the future, there are a number of issues that need to be addressed. But, it should be understood that by enhancing the reliability of our crash-test procedures, we are not guaranteed a relevant crash test procedure. Reliability is a necessary condition for a relevant crash test procedure, but it is not a sufficient condition.

Test validity might be defined as "the degree to which a test actually measures that which it purports to measure." If a crash test procedure is valid, that procedure is somehow predictive of what should occur under similar circumstances in real-world crashes. When we run crash tests according to the procedures outlined in NCHRP 350, we assume that there is some correlation between test outcomes and the injuries sustained in similar, real-world crashes. The question might then be asked: are the crash testing procedures and acceptance criteria established in NCHRP valid? If they are valid, then appurtenances installed pursuant to those standards should promote reductions in injuries in real-world collisions.

There is substantial "face validity" for the testing procedures in NCHRP 350. The test protocols and acceptance criteria appear to be predictive of the outcomes of real-world crashes, but, the protocols and test criteria of NCHRP 350 have not been tied to real-world crash experiences. In fact, it would be very difficult to experimentally correlate the procedures and acceptance criteria established in NCHRP 350 with the outcomes of real-world crashes for several reasons, including those shown in table 2.

The dependent variables in crash tests include a number of qualitative and quantitative measures (e.g., occupant impact velocity, ride down acceleration, vehicle overturn, occupant compartment penetration, etc.) that, in essence, are surrogates for injury—the dependent variable of interest in real-world crashes. Undoubtedly, there is some correlation between the crash test criteria established in NCHRP 350 and real-world injury, but how good is that correlation? The answer is we do not really know. If the occupant impact velocity criterion in NCHRP 350 were lowered, and if appurtenances were designed to meet this more rigorous criterion, to what degree would crashes on this "new and improved" appurtenance result in fewer and less severe injuries? If the ride down acceleration criterion in NCHRP 350 were lowered, by what degree would new appurtenances manufactured to meet this enhanced requirement also serve to reduce the frequency and severity of injuries to motorists striking the new appurtenance? While we do not now have good answers to the questions just posed, there is little likelihood that we will ever be able to carry out the kinds of experimental studies that might be able to answer these questions.

Many factors affect the severities of injuries sustained in motor vehicle crashes involving roadside appurtenances: the type of crash (i.e., was the vehicle yawing or pitching when it struck the appurtenance, was the impact with the front quarter panel, the front, or the side of the vehicle), the type of vehicle in which the motorists were riding, their ages, whether or not they were belted, whether or not their vehicle had an air bag that deployed during the crash, and so on. Because so many factors (independent of the quality of the appurtenance) affect crash injury, it would be difficult to tease out these extraneous factors and express crash injury as a function of the provisions in NCHRP 350, or any other standard or guideline, for that matter.

Further complicating our problem is the fact that many appurtenances that are "unacceptable" under NCHRP 350 are not being fielded. Ethics aside (which is not being proposed), if we could randomly install some new appurtenances (e.g., crash cushions) that were "woefully inadequate," "almost acceptable" and "completely acceptable" under the provisions of NCHRP 350, we might experimentally discern a correlation between NCHRP 350 provisions and crash injury. But with a truncated range of appurtenances making their way into the highway environment on a non-random basis, correlations between NCHRP 350 provisions and crash injury have been, and will continue to be, difficult to establish experimentally.

Which brings us to the issue of accident data. Police-level accident data are highly subjective, often erroneous, and typically lacking in details on appurtenance crashes (e.g., was the guardrail struck on end or at some point throughout the length of need). Accident data bases that have been assembled through in-depth investigations are somewhat more reliable and complete, but the numbers of available crashes in these data bases (for a given appurtenance) are relatively few. In short, the crash data that might be needed to experimentally validate crash testing procedures are not available.

Even if crash data were available and sufficiently reliable for purposes of evaluating the provisions of a set of crash testing procedures, it should be recognized that "reporting thresholds" can seriously impact the evaluation of crash-phase safety devices. It has been known for some time, that as reporting thresholds are raised, the estimated effectiveness of crash-phase safety devices is reduced (see, for example, Griffin 1990). "Successes"—non-injury crashes that do not make their way into the data base—are not accounted for in estimates of effectiveness.

Finally, to introduce one last impediment to the experimental evaluation of the validity of crash test procedures, it should be recognized that many appurtenances are not installed and maintained "as tested." If a longitudinal barrier that has been found acceptable under the provisions of NCHRP 350 is installed on a down slope on the outside of a horizontal curve in weak soil, and then found to be inadequate (i.e., vehicles striking the barrier are going over or through it with consequential injuries to vehicle occupants), should we conclude that NCHRP 350 is invalid? No. The real-world barrier installation is so discrepant from the staged installation for crash testing that any real-world injury outcomes sustained in impacts with this barrier do not speak to the validity of the test procedures. By the same token, barriers that are ill maintained may deviate so greatly from the articles that were tested that injuries sustained on the ill-maintained article do not imply invalidity in the testing procedures.

In lieu of an experimental assessment of the validity of a set of crash testing procedures, we might instead clinically review real-world operational data and real-world crash data to shed some light on crash testing procedures. Such clinical analyses might be particularly useful in considering: