U.S. Department of Transportation
Federal Highway Administration
1200 New Jersey Avenue, SE
Washington, DC 20590
202-366-4000


Skip to content
Facebook iconYouTube iconTwitter iconFlickr iconLinkedInInstagram

Federal Highway Administration Research and Technology
Coordinating, Developing, and Delivering Highway Transportation Innovations

Report
This report is an archived publication and may contain dated technical, contact, and link information
Publication Number: FHWA-RD-95-197
Date: December 1996

Development of Human Factors Guidelines for Advanced Traveler Information Systems and Commercial Vehicle Operations: Comparable Systems Analysis

 

CHAPTER 5. THE NAVMATE SYSTEM

 

GENERAL SYSTEM DESCRIPTION AND OBJECTIVES

USER INTERFACE

DESIGN GUIDELINES USED

LESSONS LEARNED

 

GENERAL SYSTEM DESCRIPTION AND OBJECTIVES

Navmate is an in–vehicle ATIS prototype system designed by Zexel USA Corporation to allow drivers the capability of reaching a specific destination from any place of origin. The information obtained for this system originated from three sources: a one–page brochure of Navmate given to drivers; an interview with a member of the design team; and a modest usability study with eight drivers, including three authors of this report. The Navmate analysis relies considerably on the heuristic analyses performed by the authors while performing probe tasks and driving tasks. Verbal protocols from the analysts and other subjects were recorded as the tasks were performed and subsequently transcribed.

Navmate is a route planning and navigation system that is currently undergoing fieldtesting by Zexel and Avis in San Jose, California. The Navmate is a single–destination system that utilizes a combination of vehicle location techniques and provides the driver with turn–by–turn route guidance information en route to the selected destination. Drivers can select from a number of destinations by proceeding through various text–based menu screens and selecting a point of interest or a particular intersection from a database. A driver also may enter a street address as a destination. Once a destination is entered, Navmate calculates the fastest route, taking into consideration distance, posted speed limits, number of intersections, and left/right turns. The route is displayed on a digital map showing varying levels of detail contingent on the driver–selected map scale. The smallest scale displays one complex intersection (e.g., a cloverleaf freeway interchange) in its entirety on the screen.

Table 4 denotes the functional characteristics of the Navmate system as compared to ATIS systems. Currently, the Navmate system is mostly an IRANS system that primarily provides navigation and route guidance information based on position information and map databases. However, the hardware configuration of the system renders it capable of adapting to future software systems that can provide several other ATIS functions. Plans are in progress to develop a version of the Navmate system capable of two–way communication.

 

Table 4. Comparison of Navmate functions with those from ATIS/CVO systems.

Subsystem Function Navmate
  Trip Planning  
  Multi-Mode Travel Coordination  
  Pre–Drive Route and Destination Selection *
  Dynamic Route Selection *
IRANS Route Navigation *
  Route Guidance *
  Automated Toll Collection  
  Route Scheduling (CVO-Specific)  
  Computer-Aided Dispatch (CVO-Specific)  
  Broadcast Services/Attractions  
IMSIS Services/Attractions Directory *
  Destination Coordination  
  Message Transfer  
  Roadway SignðCGuidance  
ISIS Roadway SignðCNotification  
  Roadway SignðCRegulatory  
  Immediate Hazard Warning  
  Roadway Condition Information  
IVSAWS Automatic Aid  
  Manual Aid Request  
  Vehicle Condition Monitoring  
  Cargo and Vehicle Monitoring (CVO-Specific)  
  Fleet Resource Management  
CVO-Specific Dispatch  
  Regulatory Administration  
  Regulatory Enforcement  

 

Top

 

USER INTERFACE

General Description

The Navmate unit has a 9–cm x 6.25–cm, 16–color, LCD display that provides both text– and map–based information to the driver. The display unit can be attached to the dashboard with a special mount such that it is located to the right of the driver, near the top of the dashboard. The display unit swivels and adjusts horizontally and vertically. The current Navmate prototype, including all hardware, wiring, and antenna, is designed to be installed in a test vehicle (e.g., rental car) in less than 1 hour without making any permanent modifications to the vehicle. The display is mounted onto the dashboard using existing screws in the vehicle. Therefore, the location of the display is contingent upon the location of existing screws. The GPS unit and other hardware are mounted in the trunk. The adaptability of this quick–installation design allowed the prototype to be tested in several vehicle models.

The display unit has several discrete–setting controls that provide access to the screen displays and control cursor movement. The entire unit is approximately 14 cm wide and 14.5 cm high.

The system utilizes a combination of positioning techniques, including the Global Positioning System, dead–reckoning, and map–matching. It provides the driver with real–time vehicle location and route guidance information as the driver progresses to a selected destination.

 

Visual Information Display

When the ignition of the vehicle is in the "On" position, power is available to the Navmate. When power is available and the Navmate's "On/Off" switch is turned to "On", an introductory message is displayed. This message notifies the driver that he/she should always drive defensively and implies that the device should not hinder driving safely (figure 43).

Once the driver acknowledges the message, the main "Where do you want to go?" menu is displayed. All the menus are text–based and presented with color schemes to enhance visibility and contrast. While most of the text is white on a black background, several portions of the display (header and help boxes) make use of other color schemes on a colored background.

Message displayed when vehicle ignition is turned on

The main menu screen is accessible when the vehicle is stationary or in motion. It allows a destination to be selected. From the main menu, the driver has four options for choosing the desired destination:

  • Street Address.
  • Intersection.
  • Points of Interest.
  • Guidance History.

The desired selection is chosen by moving the cursor (highlighted box) to the item and pressing ENTER. At the bottom of the screen, an instructional line indicates the manner in which the driver can make his/her selection. Icons of arrows represent cursor controls. "Points of Interest" is the most frequently used category. The directories under "Points of Interest" included the following:

  • Airports.
  • Gas Station.
  • Hotel.
  • Restaurant.
  • Rental Car Base.
  • Sports Complex.
  • Theater.
  • Toll Booth.
  • Tourist Attraction.
  • Train Station.
  • Winery.
  • Yacht Harbor.
  • Zoo.

After a specific destination directory is selected (e.g., Restaurant), the system requests that the driver "Select the Listing Mode." The driver is capable of listing each directory alphabetically or by distance (the closest restaurant at the top of the list). When directory items are listed, the driver can scroll through the selections until the desired destination appears in the fixed highlighted box. As selections are highlighted, the direction, distance, and address (and city) of that selection are displayed at the top of the screen. Once the driver enters the desired destination (e.g., Mr. Right's Cafe), the system displays a message indicating that it is calculating the shortest route. This may take several seconds, depending on the distance to the destination.

Once the route is calculated, a map–based display is shown with the route highlighted in magenta (figure 44). All local streets are white, freeways are light blue, and the background is black. Street names are in yellow. At the bottom of the map display, the compass direction in which the vehicle is heading (e.g., NW), the GPS status icon, and the map scale icon are shown in green.

Navmate map display screen

The map displays the position of the vehicle with a triangle indicating the vehicle heading. The driver can select to have the map displayed in either the heading–up or North–up orientation. Most drivers seem to prefer the heading–up mode. The location of the destination is depicted with a solid dot.

Once the destination has been selected and the route is established, the driver is not capable of altering the route; however, he or she can select a new destination. The driver also can change the system's setup (e.g., the map scale) or switch between the map and route guidance screens.

When the vehicle begins to move, the system automatically switches from the map to the route guidance screen (figure 45). The driver does have the option, however, of selecting the map display while in motion. The route guidance display provides turn–by–turn instructions for proceeding to the destination. The guidance display indicates the name of the street for the next turn at the top of the screen in uppercase letters. Underneath is a large yellow arrow indicating the direction of the next turn. Just below the arrow, the approximate distance to the next turn in displayed in miles. A yellow bar appears at the bottom of the guidance screen when the vehicle is 0.3 km from the next turn, and shrinks as the turn is approached. To the lower right of the screen, the direction of the destination from the current position is indicated with a small arrow in white. Next to it is the "as the crow flies" distance to the destination. The same GPS status icon from the map display is shown on the lower left.

Navmate route guidance display

If the driver leaves the planned route for any reason, the system displays a message indicating that the driver has "Left the Route" and also asks whether a new (alternate) route should be calculated to the same destination from the current position. The driver acknowledges this request simply by pressing ENTER.

 

Auditory Information Display

A digitized male voice provides route guidance information to support visually displayed information at two volume levels (low or high), or it can be deactivated. When the vehicle approaches a turn (0.16 to 0.80 km away, depending on the vehicle speed), the system provides a voice announcement such as "Left Turn Ahead."

A short auditory beep is emitted when a key is pressed, providing feedback that the function is engaged. Longer beeps are emitted when the system cannot respond to the key activated. A short beep is also emitted just before each voice message. All beeps seemed identical in pitch and loudness, and intensity was not adjustable. Two short beeps are emitted just before the vehicle reaches a turn.

 

User Input (Controls)

The Navmate system has several discrete controls located below the display (figure 46). All cursor movements are controlled by activating a large, circular, four–way, rocker control. To activate "Quick Scroll," the driver has to simultaneously press the rocker key and either of the Quick Scroll buttons. Several drivers seemed to have difficulty with this function. The four–arrow control is also used to relocate the vehicle position on the display when a miscalibration occurs (this may occur due to a weak GPS signal).

The ENTER control activates the highlighted feature, and CANCEL acts as an "escape" function, allowing the driver to return to the previous screen or to abort a route calculation.

The ROUTE/MAP button allows the driver to switch back and forth between the map and guidance screens while driving. The OPTIONS key is used to change the system settings for map size (small, medium, large), map orientation (heading up, North up), and Voice Volume (high, low, off). These setup settings may be changed while the vehicle is in motion.

Navmate controls

 

Communications Systems

At the present time, Navmate is a one–way communication system (GPS receiver) and does not transmit any signals or information. The map database must be updated by physically loading new databases.

Cognitive Demands

The majority of the system navigation and destination selection functions are normally performed when the vehicle is in PARK or at zero speed. Under normal conditions of use (i.e., destination is selected while parked), very few control inputs are used while driving; thus, the cognitive demands are limited to the time–sharing requirements of glancing at the display under various traffic conditions. The map display is difficult to interpret with a quick glance and can increase attentional demands on the driver. Normally, the simpler guidance display is used while in motion. Because of the simplicity of the guidance display and the distance cuing provided by the speech system, there is no need to look at the display frequently. The two beeps at the moment of the turn reduce the need to watch the display for distance information immediately prior to the turn. In short, the cognitive demands imposed by the Navmate system are not high as long as the driver selects the destination while parked.

 

System Temporal Requirements

The display change that requires the most amount of time is the calculation of a new route. At times, this can take several seconds (up to 30 seconds) and can lead to user frustration. The map display also takes several seconds to update as the vehicle makes a turn. This is likely to occur when making a U–turn or turning around in a parking lot.

 

Top

 

DESIGN GUIDELINES USED

Human Factors Design Guidelines

The Navmate design team was composed of 15 to 20 members with backgrounds in various areas of engineering, such as hardware, software, or computer science. Human Factors engineers were not involved in the design of this system, except for a Human Factors consultant who was asked to briefly evaluate the system's interface. This input was not part of a formal evaluation and the consultant did not participate in any rapid prototyping of or experimentation on Navmate.

 

Other Guidelines

The design of the system seemed to revolve around a "passive system" that would not require input from the driver once the controls were set. The design of the displays also followed this "passive system," providing simplified turn–by–turn guidance; therefore, the driver would not be required to attend to the screen frequently or for long periods of time.

Zexel design team members recognized that rental car users are likely to be novice users of the system, and too many controls could be intimidating. The interface was designed to "look easy to use."

Cost was a constraint in the design. The availability of small, low–cost, color displays constrained the system to using fewer colors and lower resolution than desired. Since all graphics and text were drawn by Zexel, the design was only limited by the screen size, resolution, and color capability. Anti–glare devices are also expensive and Zexel opted for using gray filters rather than polarizing filters.

Available technology also impacted the design of Navmate. The infrastructure was not available to provide real–time traffic information. In the future, Zexel plans to use a cellular phone to provide automated destination coordination. Route searches also were not as exhaustive as desired. The present Navmate route search is limited to reduce the amount of time needed to make a larger search. It was anticipated that rental car drivers would not be patient during long search times; therefore, it is possible that the best route (e.g., freeway around a city) is not selected.

Other technology constraints involved the inaccuracy of GPS systems, which can introduce a position error as large as 100 to 300 m. Furthermore, the GPS signal is lost in many situations (e.g., parking garages, tunnels, tall buildings). To compensate for this error, several on–board systems were integrated. Data from the car's odometer and from a gyrometer in the Navmate system are used with dead–reckoning and map–matching technology to attempt to reduce position error.

Updates and modifications to the present Navmate prototype will be made contingent upon the feedback from the current field tests. Customer Satisfaction Surveys completed by rental car users are expected to provide valuable information, including ease of use, ability to reach destinations, ease of entering destinations, and overall system performance. Drivers also are asked the price that they would pay to have the Navmate system installed as optional equipment in a new vehicle (i.e., $999 to $1,999).

 

Top

 

LESSONS LEARNED

[NM 01] SIMPLE USER ACCESS TO LARGE DATABASES IS ESSENTIAL

  • The motorist information for the Bay Area that was contained in the Navmate database was large. The method used for scrolling through long lists was awkward, sometimes requiring the operation of two buttons simultaneously.

  • More work needs to be done on enabling users to access large databases quickly. Several drivers required two hands to use the Speed Scroll button in conjunction with the large, four–position button. Subsequent design modification using button–press pressure for slow/fast scroll is better, but still awkward.

[NM 02] DETAILED ACCURACY OF MAP DATABASE MUST BE MAINTAINED

  • Map databases for ATIS/CVO systems must account for one–way streets and two–way streets with a center divider. Errors of this type can cause incorrect guidance. The consequences of this error range from inconvenient to dangerous (if the driver follows the incorrect guidance).

  • Updating map databases will be a challenge that must be met to provide the foundation for accurate ATIS/CVO systems.

[NM 03] ENTIRE ROUTE PREVIEW IS DESIRABLE

  • Users want the capability of previewing an entire route on the digital map at a selectable scale.

  • Providing this option during the destination selection process (pre–drive mode) should be considered.

[NM 04] DRIVERS WOULD LIKE TO BE ABLE TO REJECT UNVIABLE ROUTES

  • Several drivers commented that ATIS systems should allow drivers to reject an unviable route (e.g., due to inaccuracy in the map database) or request a different route than the one provided (e.g., one that excludes a particular street).

  • In–vehicle navigation systems should provide users with the ability to request an alternate route to the one provided.

[NM 05] AUTOMATED STREET SEARCH WAS VALUED WHEN ENTERING ADDRESS DESTINATION

  • When entering an intersection as a destination, the system automatically limited the second street options to only those streets that intersect the first street. This was viewed very favorably by the drivers. Several drivers were surprised that this feature existed and felt it saved much time in scrolling through the set of possible intersecting streets.

  • ATIS database access time can be shortened by this type of "shortcut" and similar aides should be integrated into each of the destination selection searches.

[NM 06] BUTTON FEEDBACK CAN IMPROVE PERFORMANCE

  • Tactile feedback from input buttons was marginal. The tone accompanying the registration of a button press was useful feedback, but several drivers felt there were too many tones. Buttons with better tactile feedback would be an improvement.

  • Adequate tactile feedback from ATIS system controls should be provided.

[NM 07] DRIVERS SHOULD BE NOTIFIED WHEN VEHICLE POSITION CALIBRATION IS REQUIRED

  • The capability of adjusting the position of the symbol of the vehicle to the correct position in the display was good. However, few of the drivers realized that this could be done, and none made use of the feature without prompting from the experimenter. The lack of use of this manual correction capability indicates that many drivers did not recognize when the system was generating guidance based on an erroneous location.

  • An intelligent system to estimate probable error and to notify the driver would be useful.

[NM 08] ROUTE DEPARTURE NOTIFICATION SHOULD BE PROVIDED VERBALLY AND VISUALLY

  • The visual display indicating that the driver is no longer following the route was sometimes not detected.

  • An auditory indication could be helpful, but would be tolerated only if the high false alarm rate of this indication was reduced.

[NM 09] SOME DISPLAY SYMBOLOGY WAS CONSISTENTLY IGNORED BY DRIVERS

  • Some of the display symbology was irrelevant to the drivers, such as the GPS status symbol. Other symbology of low relevance to the drivers included the direction of the destination, the direction of north, and the "as the crow flies" distance to the destination.

  • Because such items have low relevance, they contribute to screen clutter and should be eliminated.

[NM 10] COMPUTING/WORKING INDICATION WOULD CLARIFY SYSTEM STATUS TO DRIVERS

  • Route calculation time is dependent on destination distance and often takes several seconds. There is no display indication that the system is actually active and computing a route. Drivers felt something was wrong. The lack of such an indicator was noted particularly when the length of the route was very long, i.e., 80.5 km.

  • Some indication that the computer is busy calculating a route seems to be advisable.

[NM 11] ACCURACY OF MAP DATABASE IS CRITICAL IN ESTABLISHING DRIVER CONFIDENCE

  • On some occasions, the system did not find the "closest" destination known by the driver (e.g., bank). Similarly, the system sometimes provided an outdated destination for a gas station or a store. These types of errors lead to reduced driver confidence in the system.

  • The underlying problem in both these instances is outdated map database information that should be minimized in ATIS system design.

[NM 12] DISPLAY MODE CHANGES SHOULD BE UNDER USER CONTROL

  • The device changed to guidance mode from text mode automatically. The amount of time the drivers had to review the route in text mode was insufficient.

  • In general, users like to have control over display options.

[NM 13] ROUTE GUIDANCE MODE PREFERRED FOR DRIVING

  • In most cases, the drivers selected the guidance mode rather than the map mode while driving. The only time the map mode was preferred was when driving down a freeway and monitoring for the desired off–ramp.

[NM 14] COLOR CODING OF ROUTE WAS HELPFUL

  • Showing the recommended route in a different color than the streets was considered helpful by the drivers.

  • Color coding allowed drivers to easily perceive the route to be followed.

[NM 15] DISPLAY MOUNTING LOCATION WAS ADEQUATE

  • The mounting position did not obscure the roadway for any of the drivers, nor did it obscure or limit access to any controls.

  • The ideal mounting position would appear to be as high as possible without obscuring any of the roadway. This position, obviously, depends on individual measurements.

[NM 16] LEARNING REQUIREMENTS WERE FEW

  • All drivers estimated that it would take less than 1/2 day to learn to operate the system well, and most indicated that the time required would be between 1 and 2 hours.

  • ATIS systems need to be designed so that the driver does not have to rely on instruction manuals and reference cards to operate the system.

[NM 17] TRAINING AIDES SHOULD BE USED AS A REFERENCE INSTEAD

  • None of the drivers went through the example in the one–page brochure. Most drivers felt the brochure was too cluttered. However, a number of the drivers referred back to the card when they were unable to quickly perform tasks.

  • Brochure design should take into consideration use as a reference tool.

[NM 18] DIGITAL SPEECH SYSTEM WAS INTELLIGIBLE

  • Use of digital voice is often considered questionable due to its limitations. Drivers felt that the digital voice was intelligible to all of the drivers in this study.

  • Future use of digital speech systems in ATIS applications is feasible.

[NM 19] PREFERRED MAP ORIENTATION WAS HEADING UP

  • The drivers generally preferred the map in heading–up, rather than north–up orientation. The one exception was an individual with extensive map reading experience.

  • ATIS systems should always provide the heading–up orientation as an option, or as the default (if only one orientation is possible).

[NM 20] DRIVER PERCEPTION OF INTERFACE MAY IMPACT USE

  • The device was not perceived as intimidating. Consequently, users were open to using and exploring the system's capabilities. In part, this seems to be due to the small number of user controls (buttons). It may be that the willingness to "play" with the system facilitates rapid learning of the system's features.

  • Simple interfaces with few controls seem to foster system use and acceptability.

[NM 21] MINIMAL DISPLAY CLUTTER ENHANCES PERFORMANCE

  • The earlier prototype of Navmate had more information on each display. However, Zexel design team members seemed to think that clutter was a problem that would lead to a long interpretation time. Therefore, several items were eliminated from the displays.

  • The current display clutter seemed acceptable, although it could be reduced further (see NM 09).

[NM 22] ATIS SYSTEM USE AT NIGHT WAS EXTREMELY VALUED

  • Navmate was tested by only two drivers in night driving situations. They found the system to be even more useful at night than it was during the day because way finding is difficult under conditions of reduced visibility.

  • If in–vehicle navigation systems have more utility at night, the interface should be designed to meet the appropriate environmental conditions.

[NM 23] REDUCTION OF DISPLAY GLARE WAS SUCCESSFUL

  • Gray filters seemed to be a cost–effective way to reduce display glare since drivers did not perceive glare as a significant problem. However, there were occasions that glare was apparent when direct sunlight shined on the display.

  • Glare–reducing filters may be sufficient to prevent display reflections.

[NM 24] DIRECT DISPLAY GLARE AT NIGHT WAS NOT A PROBLEM

  • Glare from the headlights of other vehicles was not a problem.

  • The screen was low enough that lights from following vehicles did not strike the screen.

[NM 25] COMMENTS ON VOICE MESSAGE TIMING VARIED

  • Drivers seemed to question when it would be most useful for voice messages to convey a turn. The ideal time from the onset of the voice to the actual turn needs to be researched. Opinions ranged from there not being enough time to the amount of time being about right. There were no reports of the turn alert occurring too early.

[NM 26] SUPPLEMENTAL VOICE ROUTING MESSAGES HELPFUL

  • The speech output worked well in conjunction with the visual display of route guidance to cue the driver as to the relative distance and timing of upcoming turns.

  • Visual and auditory displays should be combined in route guidance.

[NM 27] DRIVER–REQUESTED ROUTE INFORMATION

  • Drivers felt route information on estimated time of arrival (ETA), elapsed time, and ground distance to destination should be displayed at all times, or at least be easily selectable.

[NM 28] DRIVER–REQUESTED TRAFFIC CONTROL INFORMATION

  • Speed limit information is available to the Navmate, and it is used in selecting the "fastest" route, according to a design team member. However, this information is not displayed to the driver, and some drivers expressed interest in the availability of a speed limit display.

[NM 29] COMPUTING AN ALTERNATE ROUTE TO A GIVEN DESTINATION

  • Generating a new route to the same destination when you have left the route is a single button press. The amount of time required is comparable to turning on the air conditioner or changing channels on the radio.

  • Allowing re–computation of a new route to the same destination on the fly with a single button press works out well with this system. This is preferable to having to stop and perform the re–computation in the pre–drive mode.

[NM 30] SELECTING NEW DESTINATION AND ROUTE WHILE EN ROUTE

  • In the situation where there is more than one person in the car, there may not be a safety problem with having the passenger operate the system. In such a situation, a positive lock–out (preventing the selection of destinations while in motion) would be perceived as a drawback. On the other hand, when a driver is alone in the vehicle, entering a destination would require considerable time, with attention devoted to the device instead of to driving the vehicle.

  • This is a potential safety issue that needs to be addressed.

[NM 31] TURN INDICATORS SHOULD REFLECT ROADWAY CHARACTERISTICS

  • The shape of the arrows indicating a turn should represent the size and shape of the physical turn. The arrows used in the guidance mode generally provided sufficient information to the drivers regarding the angular size of upcoming turns.

  • The arrows indicating whether a U–turn, a 90E (approximately) turn, or a turn less than 90E was ahead provided an adequate number of categories for the driver to plan ahead. The consensus of the drivers seemed to be that the turn arrows adequately conveyed the turn requirements.

[NM 32] CONSECUTIVE TURN INFORMATION WAS VALUABLE

  • The verbal warning that a turn was closely followed by another turn was judged favorably. This warning generally allowed drivers to position their vehicle to perform sequential maneuvers.

  • Drivers should be made aware of consecutive turns.

[NM 33] REAL–TIME DISTANCE TO A TURN WAS VALUABLE

  • Approximate real–time distance to a turn is valuable information. The distance to turn tape [a yellow bar was displayed on the bottom of the screen when the vehicle was within 0.32 km of the turn, according to the brochure] provided a useful cue to the driver that a turn was to be performed in the immediate future.

  • Real–time distance should be made available to drivers, possibly as a display option.

[NM 34] FALSE ROUTE DEPARTURE NOTIFICATION SHOULD BE MINIMIZED

  • There was an annoyingly high frequency of false "you have left the route" messages displayed within the first minute of departing (typically in situations when buildings or structures obscured satellite signal strength). Most drivers adopted the strategy of ignoring the message and having a new route computed while driving (one button push).

  • Route departure notification messages should be accurate and should not be dependent on signal strength. A different message could be displayed when the signal is weakened or blocked.

 

Top

 

FHWA-RD-95-197

Back | Table of Contents | Next

Federal Highway Administration | 1200 New Jersey Avenue, SE | Washington, DC 20590 | 202-366-4000
Turner-Fairbank Highway Research Center | 6300 Georgetown Pike | McLean, VA | 22101