Skip to contentUnited States Department of Transportation - Federal Highway Administration FHWA Home
Research Home
Public Roads
Featuring developments in Federal highway policies, programs, and research and technology.
This magazine is an archived publication and may contain dated technical, contact, and link information.
Federal Highway Administration > Publications > Public Roads > Vol. 71 · No. 3 > A New Look at Sensors

Nov/Dec 2007
Vol. 71 · No. 3

Publication Number: FHWA-HRT-08-001

A New Look at Sensors

by David Gibson, Milton K. "Pete" Mills, and Lawrence A. Klein

FHWA's updated Traffic Detector Handbook describes in-roadway and over-roadway technologies for vehicle detection as key to ITS deployment.

Personnel from FHWA and the Virginia Department of Transportation are calibrating a video image processor (VIP) installation. VIPs are one type of over-roadway sensor.
(Above) Personnel from FHWA and the Virginia Department of Transportation are calibrating a video image processor (VIP) installation. VIPs are one type of over-roadway sensor.

With more vehicles on the roads and less available capacity, congestion has become a national problem. Part of the challenge is to manage the movement of more vehicles on existing infrastructure, which is one objective of intelligent transportation systems (ITS).

ITS applications rely on traffic flow sensors to provide vehicle detection; incident detection; ramp metering data; real-time traffic adaptive signal control; roadway volume and vehicle classification archival and planning data; and data for traveler, commercial, and emergency information services. The success of these ITS applications depends largely on the proper design, installation, and maintenance of sensor components.

Over the past 17 years, sensor manufacturers have developed new, more effective vehicle detection technologies, particularly for over-roadway sensors. During the same period, the manufacturers discontinued some over-roadway and in- roadway models. The new hardware is better able to meet the expectations of the ITS community.

To educate the transportation community on the latest sensor technologies, the Federal Highway Administration (FHWA) recently published a revised and restructured edition of the Traffic Detector Handbook (FHWA-HRT-06-108 and FHWA-HRT-06-139), a two-volume, comprehensive reference on sensors for traffic management on surface streets, arterials, and freeways. Previous editions of the handbook, published in 1985 and 1990, required updating to reflect the evolution, maturation, and state of the practice of sensor hardware and installations.

Antoinette Wilbur, former director of the FHWA Office of Operations Research and Development, writes in the revision's foreword: "The objective of the third edition of the Traffic Detector Handbook is to provide a comprehensive reference document to aid the practicing traffic engineer, planner, or technician in selecting, designing, installing, and maintaining traffic sensors for signalized intersections and freeways." She adds, "The information contained in this handbook is based on the latest research on available treatments and best practices in use by jurisdictions across the United States and elsewhere."

The revised handbook discusses selecting, configuring, installing, operating, and maintaining traffic sensors, along with new applications of sensors to advanced signal control, ramp metering, incident detection, efficient corridor operation, toll collection, collection of travel time and other data, priority vehicle and pedestrian detection, vehicle/driver safety, and other ITS functions. The enhanced descriptions of ITS applications and the other improvements discussed in the revised edition can help departments of transportation (DOTs) ensure long-term savings of public funds.

Traffic Detector Handbook, Volume I. Traffic Detector Handbook, Volume II.
Traffic Detector Handbook, Volumes I and II.

Types of Sensors

As defined in the handbook, "A traffic flow sensor is a device that indicates the presence or passage of vehicles and provides data or information that supports traffic management applications such as signal control, freeway mainline and ramp control, incident detection, and gathering of vehicle volume and classification data to meet State and Federal reporting requirements."

In-roadway sensors are embedded in the pavement or the subgrade, or they are taped or otherwise attached to the surface of the roadway. Over-roadway sensors are mounted above the roadway or alongside it.

One of the main types of in- roadway sensors is the inductive- loop detector, which consists of loops of wire embedded into sawcuts in the road pavement. A conductive metal object, such as a vehicle passing over or stopped within the sensor's detection area, decreases the loop's inductance (an electrical property), producing an electrical signal that is transmitted through a curbside junction box (a "pull box") to an electronics unit housed in a controller cabinet. The electronics unit analyzes the signal, interpreting it as the presence or passage of a vehicle, and sends an appropriate call to the controller.

According to the revised handbook, "Today, the inductive-loop detector is, by far, the most widely used sensor in modern traffic control systems." The handbook clarifies the calculations required to design properly functioning inductive-loop detector systems for intersection and highway applications.

This electronics unit for an inductive-loop sensor is capable of identifying more than 20 vehicle classes. Photo courtesy of Reno A&E, Reno, NV. This electronics unit for an inductive-loop sensor is capable of identifying more than 20 vehicle classes.

Other in-roadway sensors include magnetic detectors and magnetometers, which can be placed underneath a roadway or bridge. A magnetic detector senses changes in the Earth's magnetic field caused by passage of a nearby vehicle that contains ferrous material. A magnetometer measures the difference in the Earth's magnetic field caused by the passage or presence of a vehicle. Its ability to function as a presence sensor enables it to detect stopped vehicles. Because both of these sensors are passive devices, they do not transmit energy. Therefore, a portion of the vehicle must pass over the sensor for it to be detected. A magnetometer can detect two vehicles separated by as little as 0.3 meter (1.0 foot). This potentially makes the magnetometer as accurate as — or even better than — the inductive-loop detector at counting vehicles.

Examples of over-roadway sensors include video image processors, which use cameras mounted on tall poles adjacent to the roadway or on traffic signal mast arms over the road; microwave radar, laser radar, ultrasonic, and passive infrared sensors installed either alongside or above the road; and acoustic sensors installed alongside the road. The required mounting configuration is a function of the intended application. Modern over-roadway sensors provide a viable alternative to inductive-loop detectors.

Key changes in the new handbook include descriptions of enhanced infrared and microwave sensors, video image processors, and magnetometer sensors.

Sensor Applications And Functions

This wireless magnetometer is a type of in-roadway sensor that also can be placed underneath a roadway or bridge. Photo: Sensys Networks, Inc., Berkeley, CA.
This wireless magnetometer is a type of in-roadway sensor that also can be placed underneath a roadway or bridge.

Sensor applications for traffic control and management continue to evolve. Originally used for signalized intersection control, sensors now supply real-time data for traffic adaptive signal control, mitigation of recurring and nonrecurring congestion on freeways, and gathering of volume and vehicle classification information for road use and planning purposes, among other applications.

The technologies discussed in the handbook are time tested for traffic management applications, although some might not provide the data required for a specific use. Some technologies, such as video image processing, microwave and laser radars, and inductive-loop detectors, continue to evolve by adding capabilities that measure additional traffic parameters, such as vehicle length, classification, or acceleration; track vehicles; improve spatial resolution; or link data from one sensor to those from another. Combinations of technologies are being integrated into one unit by manufacturers to provide more robust data under a variety of traffic flow conditions.

Most vehicle sensors in use today monitor the movement of vehicles past a given point. The data acquired are transmitted to a signal controller, traffic counter, or other device. The controller or counter processes some data locally, while other data are transmitted to a central computer or display monitor, in the case of camera imagery, at a traffic management center.

These two technicians are installing a wireless magnetometer sensor, flush with the surface in a pavement borehole.
These two technicians are installing a wireless magnetometer sensor, flush with the surface in a pavement borehole.

Although single inductive-loop detectors transmit direct information concerning vehicle passage and presence, other traffic flow parameters such as density and speed must be derived from algorithms that interpret or analyze the measured data. When these parameters are calculated from inductive-loop data, the values might not be sufficiently accurate for some applications (such as rapidly detecting freeway incidents).

Pavement deterioration, improper installation, and weather-related effects can degrade the operation of in-roadway sensors such as inductive-loop detectors. Street and utility repair also can impair loop integrity. Thus, effective loop installation, acceptance testing, maintenance, and repair programs are required.

On the other hand, according to the handbook, "Over-roadway sensors are becoming more popular as sources of real-time data for signal control and freeway traffic management. This is because of their ability to provide multiple lane data from a single sensor, reduced maintenance and increased safety to installation personnel, data types not available from loops or magnetometers, and competitive purchase and installation costs."

The traffic flow parameters measured with over-roadway sensors satisfy the accuracy requirements of many current freeway and surface street applications, provided suitable mounting is available. In terms of operation, the mounting location must provide an unobstructed view of vehicles for optimum performance.

When a sensor is installed directly over the lane of traffic that it is intended to monitor, its view and hence its ability to collect data are not obstructed. But when a sensor is mounted on the side of a roadway and views multiple lanes of traffic at a perpendicular or oblique angle to the direction of traffic flow, tall vehicles can block its view of distant lanes, potentially causing an undercount or false average speed measurement. Thus, sensor type, mounting height and location, vehicle mix, road configuration, and sensor viewing angles must be analyzed in light of the intended application.

Inductive-Loop Configuration for Detecting Small and Large Vehicles

The revised handbook compares various inductive-loop digital electronics units in terms of sensitivity and response time. Newer units and loop configurations are capable of vehicle classification. Special configurations of inductive loops have been developed to detect axles and their relative position on a vehicle. Such systems are used at toll plazas to elicit the correct payment for the vehicle class. The data obtained are vehicle type, length, speed, acceleration, number of axles, and axle separation.

Microwave radar sensors such as this one transmit a frequency modulated waveform, which supports measurement of the distance between the sensor and a vehicle. This type of microwave sensor can detect both stopped and moving vehicles in multiple lanes or in multiple areas in a single lane.
Microwave radar sensors such as this one transmit a frequency modulated waveform, which supports measurement of the distance between the sensor and a vehicle. This type of microwave sensor can detect both stopped and moving vehicles in multiple lanes or in multiple areas in a single lane.

The revised handbook also contains improved explanations of inductance calculation for the detection of small and large vehicles. One of the inherent problems associated with detection of bicycles is assuring that the rider will pedal within the loop's detection zone. When a bicycle or motorcycle travels along a loop wire, eddy currents are induced in the conducting wheel rims and frame. When the cycle is directly over the loop wire, coupling between the inductive loop and the cycle is maximized — hence detection of the cycle. Loop system sensitivity is defined as the smallest change of inductance at the electronics unit terminals that will cause the controller to activate. Many States specify that the electronics unit must respond to a 0.02 percent change in inductance.

Selection of Sensors

Traffic managers selecting a sensor should consider the intended application, ease of installation and maintenance, and design requirements.

Choosing a sensor for a specific application depends on data parameters, data accuracy, spatial resolution, detection area, appropriate data transmission media, location-specific installation requirements, initial cost, and acceptability of the maintenance burden that the sensor will impose. Traffic managers should assess these criteria, separately and in combination, as part of the selection process.

This microwave radar sensor is mounted on the same pole as a camera and a camera lowering device.
This microwave radar sensor is mounted on the same pole as a camera and a camera lowering device.

Installing and maintaining in- roadway sensors such as inductive- loop and magnetic sensors can disrupt traffic and pose a safety risk to the installers. But traffic managers continue to use in-roadway sensors for a number of reasons, including (1) aesthetic considerations, (2) integration with axle counting and weigh-in-motion applications requiring sensors under or on the road surface, (3) cost and safety issues associated with mounting over-roadway sensors where existing structures are unavailable, and (4) policies that prohibit over-roadway sensors in certain locations. Newly and properly installed inductive-loop detectors also can provide more accurate data than many over-roadway sensors when they are coupled with the advanced electronics units available from several manufacturers.

As for maintenance, the passage-detecting magnetic detector, despite limited applications, has managed to retain some popularity largely because of its ruggedness and long life with minimum upkeep.

Sensor technology and operating theory indicate that the principal in-roadway sensors (inductive-loop, presence-detecting magnetometers, and passage-detecting magnetic detectors) are suitable for some applications but unsuitable for others. For example, magnetic detectors generally cannot be used for vehicle presence detection.

In terms of ease of installation, the handbook notes: "Today, agencies often look favorably on eliminating a sawcut or replacing it with a drilled hole. The pervasiveness of deteriorating pavements has produced more interest in installing preformed loops, microloops, or pavement slabs with sensors already in place." The same concerns often lead to the selection of above-roadway sensors.

Finally, sensor selection for vehicle detection at intersections is a function of the types of timing intervals generated by the controller and the corresponding data needed to compute the intervals. Therefore, the timing interval types should be selected early in the design process.

Additional Advantages and Disadvantages Affecting Selection

The handbook compares the capabilities, strengths, and weaknesses of current sensor technologies in terms of installation, parameters measured, and performance in inclement weather and under variable lighting and changeable traffic flow.

This video image processor, a type of over-roadway sensor, can provide traffic flow data across several lanes and in multiple areas in one lane. Photo: Econolite, Anaheim, CA.
This video image processor, a type of over-roadway sensor, can provide traffic flow data across several lanes and in multiple areas in one lane.

The good performance of in-roadway sensors such as inductive loops, magnetic detectors, and magnetometers is due partly to their close proximity to the vehicles passing over them. Another advantage is that they are insensitive to inclement weather such as rain, fog, and snow. Their main disadvantage is their in-roadway installation, which necessitates a pavement cut. Also, inductive-loop detectors are not always appropriate for some traffic signalization applications. For example, long loops are not suitable for detecting oversaturated flow or long queues of vehicles.

Because a magnetometer's probes are buried in a drilled hole below the surface, the devices are especially useful in the Northeastern States and other cold regions, where pavement deteriorates more rapidly due to thermal expansion and contraction and suffers damage from snow-removal equipment. Another advantage is that magnetometer probes and lead-in wires tend to survive in crumbly pavements longer than ordinary loops.

Unlike inductive-loop detectors, magnetometers usually will operate on bridge decks where uncoated steel is present and cutting the deck pavement for loop installation is not permitted. Another benefit is that they require fewer linear feet of sawcut. Modern magnetometers are built with both horizontal and vertical axis sensors. Therefore, they can operate virtually anywhere.

The over-roadway laser radar sensor transmits energy in the near infrared spectrum, which is just above the visible wavelength spectrum. An advantage of laser radar is that it transmits multiple beams for accurate measurement of vehicle position, speed, and class. Up to 11 standard and 20 user-defined classes can be reported. But its operation can be affected by fog or blowing snow when such conditions restrict visibility to less than the distance from the sensor to the road, typically 6 meters (20 feet). Another disadvantage is that its installation and maintenance, including periodic lens cleaning, require lane closure when it is mounted above the road.

Microwave radar sensors that transmit a frequency-modulated waveform, which supports measurement of the distance between the sensor and a vehicle, can detect stopped and moving vehicles in several lanes when mounted alongside the road. When mounted above a particular lane, they can detect vehicles in multiple areas in a single lane. But microwave radar sensors that transmit a continuous wave signal (one that is constant in frequency) cannot detect stopped vehicles and usually are limited to monitoring one lane of traffic, so traffic managers should consider these limitations when selecting a sensor for the chosen application. Data supplied by presence-detecting microwave radars include volume, lane occupancy, speed, and vehicle class by length.

Video image processors (VIPs) typically consist of one or more cameras, a computer for digitizing and analyzing the imagery, and software for interpreting the images and converting them into traffic flow data. Black-and-white image analysis is performed by algorithms that examine the variation of gray levels in groups of pixels (picture elements) contained in the video frames. A VIP system can provide traffic flow data across several lanes and in multiple areas in one lane. VIPs can classify vehicles by their length and report vehicle presence, volume, lane occupancy, and speed for each class and lane. VIPs also can register vehicle turning movements and lane changes.

But VIPs require line-of-sight views of the areas they monitor and are susceptible to inclement weather. Installation and maintenance, including periodic lens cleaning, require lane closure when a camera is mounted over the roadway. Performance is affected by vehicle shadows; the day-to-night transition; sun glint; vehicle/road contrast; and water, salt grime, icicles, and cobwebs on the camera lens. Reliable nighttime signal actuation requires street lighting and a camera mounting height of 9 to 15 meters (30 to 50 feet) in a side-mounting configuration for optimum presence detection and speed measurement. Some models are susceptible to camera motion caused by strong winds and vibration of the mounting structure.

The handbook states, "What seems to be clear is that agencies [that are] contemplating the use of video detection should approach it carefully as there are many pitfalls...It seems clear that it is also important to make sure a vendor is selected that can provide the latest improvements in video detection technology." On the other hand, the handbook adds that VIPs generally are cost effective when many detection zones or specialized data are required.

Finally, the handbook compares sensors in terms of costs for purchasing the hardware and software, plus installation, maintenance, and repair — all costs that should be factored into the selection decision. Installation costs, for example, include those for the technicians who prepare the road surface or subsurface (for inductive-loop or other surface or subsurface sensors), install the sensor and mounting structure (if one is required for over-roadway sensors), purchase and install the conduit, close traffic lanes, divert traffic, provide safety measures where required, and verify proper functioning of the device after installation is complete.

This VIP has an integrated camera and processor in one unit. VIPs can classify vehicles by their length and report vehicle presence, volume, lane occupancy, and speed for each class and lane.
This VIP has an integrated camera and processor in one unit. VIPs can classify vehicles by their length and report vehicle presence, volume, lane occupancy, and speed for each class and lane.

Installation: Best Current Practices

The mechanical operations associated with sensor installation often present challenges. For example, installing an inductive-loop detector requires approved procedures for cutting a slot in the pavement, laying turns of wire in the slot, twisting the wires, covering them with sealant, removing excess sealant, splicing the wire to the cable, and connecting the cable to the electronics unit in the controller cabinet.

For over-roadway sensors, analogous tasks and issues arise, such as installation of mounting structures, power and data cables, sensor alignment, and calibration verification. Over-roadway sensors require sign bridges, mast arms, poles, or similar overhead structures for mounting. If such a structure is not already in place, it must be installed to support the sensors.

Shown here is a VIP with integrated camera and processor in one unit. VIPs also can register vehicle turning movements and lane changes. Photo: Image Sensing Systems, St. Paul, MN.
Shown here is a VIP with integrated camera and processor in one unit. VIPs also can register vehicle turning movements and lane changes.

An over-roadway sensor's field of view, that is, the area of the roadway within which a vehicle is detected and data are collected, is a function of sensor mounting height and aperture size, offset of the mounting location from the lanes to be monitored, elevation changes and curves in the road, and objects that can block the view. These factors must be taken into account when installing this kind of sensor.

Installation techniques and theories vary widely among traffic agencies. As noted in the handbook, "Procedures developed over time frequently become outdated or are no longer effective; yet there is often great resistance to change. In many cases, contractors perform sensor installation using proprietary shortcuts (and shortcomings)." The handbook notes, "Improper or sloppy installation causes many of the sensor failures and signal malfunctions that are observed."

Key changes in the revised handbook regarding installation include updates to the procedures for installing inductive loops and inclusion of new material that describes installation of above-roadway sensors. Because of the failures attributed to moisture or breaks in wire, the trend is to encase and seal the loop wires in a protective covering prior to sealing the sawcut. Some agencies choose to prewind the wires or preform the loops in the shop to ensure the proper number of turns and reduce installation time on the roadway. In addition, many inductive-loop detectors now are built into the pavement during construction of a new roadway or during repaving.

This rack-mount VIP uses a modular approach that enables the system to be configured for specific applications that can vary the numbers of video inputs, inputs from external systems and outputs to support complex uses, and Internet browser access to video imagery. Photo: Iteris, Inc., Santa Ana, CA.
This rack-mount VIP uses a modular approach that enables the system to be configured for specific applications that can vary the numbers of video inputs, inputs from external systems and outputs to support complex uses, and Internet browser access to video imagery.

Maintenance and Troubleshooting

The handbook notes, "Use of appropriate sensor installation techniques and specification of suitable materials and products will minimize maintenance and other life-cycle costs. However, even with superior design and installation, proper and regularly scheduled sensor maintenance is critical to effective and prolonged operation of traffic signal control systems and freeway surveillance and management systems."

Many factors such as inadequate budget and staffing deficiencies can contribute to lack of maintenance. "Budgetary problems, which continue to plague traffic agencies, have resulted in a cost consciousness that frequently focuses only on initial cost, rather than on lifetime cost," according to the handbook. "Consequently, less expensive products, materials, and processes are used in the original installation because of their lower initial cost."

Maintenance issues associated with inductive-loop detectors have changed considerably over the years. For example, the inductive loop's electronics unit, which formerly accounted for a considerable portion of sensor malfunctions, has matured to the point where many currently available digital models seldom experience failures. Recognizing the heavy demand on maintenance dollars, some manufacturers added circuitry that reduces the frequency of trouble calls to reset units attached to faulty loops.

Maintenance and life-cycle costs may be determined, in part, by published values of the mean time between failures. Some over-roadway sensors are designed to operate for 35,000 to 90,000 hours before a potential failure. The effects of lightning strikes and other natural or human-induced failure modes are not included in this number. Over a 10-year period, maintenance and replacement costs for these devices can be significantly less than for inductive loops, especially if commercial vehicle loads, poor subsoil, inclement weather, and utility improvements frequently require road resurfacing and loop replacement.

Technicians reposition, recalibrate, and clean a video camera used as a traffic sensor. The video camera is mounted on a vertical elevated pole attached to a horizontal signal head mast arm. The elevated pole is visible between the two buckets the technicians are standing in.
Technicians reposition, recalibrate, and clean a video camera used as a traffic sensor. The video camera is mounted on a vertical elevated pole attached to a horizontal signal head mast arm. The elevated pole is visible between the two buckets the technicians are standing in.

A 10-year study of inductive-loop maintenance costs in Houston, TX, found as few as 42 failures and as many as 341 failures per year in the 600 to 1,000 intersections maintained during the 1989-1998 study period. The calculated loop replacement costs per intersection varied from $107 to $628. Actual costs per intersection are probably higher because the calculation assumes all intersections had loops (some were not actuated and hence did not use loops), 100 percent of loop failures were discovered (some were not), and no maintenance besides replacement was performed.

This worker is placing a tie down backer rod while installing turns of wire as part of an inductive-loop sensor.
This worker is placing a tie down backer rod while installing turns of wire as part of an inductive-loop sensor.

Similarly, in a summary of maintenance costs for four VIP systems used by the Road Commission for Oakland County, MI, monthly camera maintenance averaged $5.05, and monthly processor maintenance averaged $26.71 from 1995 through 1998. A total of 692 cameras and 194 controllers were included in the study. Costs included labor; fringe benefits; and truck, lift, and radio equipment.

Last Word

The world of traffic sensors is changing rapidly as manufacturers develop new technologies and retire older models. The new edition of the Traffic Detector Handbook will enable users to select specific technologies for various applications based on the sensors' capabilities, and to configure, install, and maintain the sensors to achieve an agency's goals for traffic management.

These workers are sawcutting slots in the pavement in which loop wires will be placed as part of the installation of an inductive-loop detector system. After the slots are cut, the debris is removed by power washing (see worker on left) and then blow drying the slot. This installation is being done at night to minimize traffic disruptions.
These workers are sawcutting slots in the pavement in which loop wires will be placed as part of the installation of an inductive-loop detector system. After the slots are cut, the debris is removed by power washing (see worker on left) and then blow drying the slot. This installation is being done at night to minimize traffic disruptions.

David Gibson is a highway research engineer on the Enabling Technologies Team in FHWA's Office of Operations Research and Development. He is a registered professional traffic engineer and has a master's degree in transportation from Virginia Polytechnic Institute and State University. His areas of interest include traffic sensor technology, traffic control hardware, modeling, and traffic engineering education. He worked with Milton K. "Pete" Mills on the first two editions of the Traffic Detector Handbook and the original Type 170 traffic signal controller system.

Milton K. "Pete" Mills is an electrical engineer, now retired, from the Office of Safety Research and Development at FHWA's Turner-Fairbank Highway Research Center (TFHRC). He holds a bachelor's degree in electrical engineering from North Carolina State University and a master's degree from The Catholic University of America. At TFHRC since 1968, he managed development and evaluation of systems for sensing vehicles from infrastructure and sensing infrastructure from vehicles. His current interests include sensor development and application, image processing methods, and Super Equation Shell software and numerical error propagation. Prior to joining TFHRC, he tested and evaluated aircraft and spacecraft antenna systems for the U.S. Navy and National Aeronautics and Space Administration.

Lawrence A. Klein, Ph.D., brings more than 30 years of aerospace and traffic management experience to the development of sensor and data fusion approaches for ITS and multiple sensor concepts for homeland security. He is the principal author of the third edition of the Traffic Detector Handbook; Sensor Technologies and Data Requirements for ITS, which discusses sensor applications for traffic and transportation management; Sensor and Data Fusion: A Tool for Information Assessment and Decision Making, which presents data and sensor fusion architectures and algorithms for identifying and tracking objects; and Millimeter-Wave and Infrared Multisensor Design and Signal Processing, which describes multisensor design and performance.

For more information, see the Traffic Detector Handbook, Third Edition, Volume I at www.fhwa.dot.gov/publications/research/operations/its/06108/index.cfm and Volume II at www.fhwa.dot.gov/publications/research/operations/its/06139/index.cfm. Contact David Gibson at 202-493-3271 or david.gibson@dot.gov, Milton K. "Pete" Mills at 202-244-1136 or pete.mills@erols.com, or Lawrence A. Klein at 310-541-2622 or laklein@flash.net.

ResearchFHWA
FHWA
United States Department of Transportation - Federal Highway Administration