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


Skip to content
FacebookYouTubeTwitterFlickrLinkedIn

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
Back to Publication List        
Publication Number:  FHWA-HRT-16-024     Date:  June 2016
Publication Number: FHWA-HRT-16-024
Date: June 2016

 

LTBP Program's Literature Review on Weigh-In-Motion Systems

Chapter 3. WIM Systems Overview

WIM Systems Development

There have been many initiatives contributing to the development and improvement of WIM systems. In the United States, the American Association of State Highway and Transportation Officials (AASHTO) Technology Implementation Group designated WIM as a concept of focus technology in 2004. The work conducted under FHWA’s LTPP Program resulted in the development of a field operations guide for WIM sites.(29) The guide contains recommendations for WIM system installation, calibration, operation, and data validation procedures that enable these systems to collect research quality data. Activities targeting WIMsystem development in Europe include the Weigh-In-Motion of Axles and Vehicles for Europe(WAVE) project and the European Cooperation in Science and Technology 323 project: Weigh-In-Motion of Road Vehicles.(30,31)

A significant challenge associated with WIM systems is installing sensors in the roadway pavement. This requires temporary roadway closures and pavement cuts for placing sensors.
The condition of the existing pavement at the installation site may also create challenges for installation and for obtaining reliable truck weight measurements. The pavement at the site must be sufficiently smooth for a minimum distance before and after the location of the weight sensor to minimize the influence of vehicle dynamics on the weight measurements. In practice, existing pavement sections often do not meet the minimum smoothness specifications for WIM system installations. This can require rehabilitation or replacement of the existing pavement at the WIM site to achieve the required smoothness. Maintaining the required smoothness at the WIM site throughout the lifespan of the WIM installation also poses a challenge.

The ASTM E1318 standard for WIM systems classifies WIM systems according to the following four distinct types (Type I through Type IV), depending on the application and functional performance requirements:(2)

The following discussion is limited to WIM systems for traffic data collection applications because weight enforcement applications are beyond the scope of this project. It is important to note that the functional performance requirements for Type I and Type II systems can be satisfied by most of the WIM system technologies currently available on the market, but the initial cost, installation, maintenance, calibration requirements, and long-term performance characteristics for the different available technologies can vary significantly.

Components of a WIM System

Although there are some portable WIM systems currently under development, using these devices for traffic data collection has revealed many installation and calibration challenges that make their use problematic for traffic data collection on high traffic routes or for extended durations.(32) The vast majority of traffic weight data collection has been and is currently performed using permanent WIM systems.

The major components of permanent WIM systems include various sensors embedded in the roadway surface to detect, weigh, and classify vehicles; software and electronics to control the WIM system sensors and collect, analyze, and store the sensor measurements; and communications hardware used to transmit the vehicle measurements offsite. The electronics and communications devices are usually located in a roadside cabinet adjacent to the WIM site, and the system is powered by either a direct AC power connection or by batteries charged by a solar panel array.

Weight Sensor

The weight sensor is the most fundamental and important component in the WIM system because it directly measures the force applied by the vehicles passing over the sensor. The principal weight sensor types used extensively for permanent WIM system installations include bending plate, load cell, and piezoelectric. These weight sensor types primarily differ according to their principle of operation. Each sensor type also has its own advantages and disadvantages with respect to its use for WIM systems.

Bending Plate Sensors

Bending plate weight sensors utilize strain gages that are mounted to the underside of high‑strength, rectangular steel plates called weighpads. The strain gages are wired in a Wheatstone bridge circuit configuration, and when a wheel passes over the weighpad, the WIM system software uses the measured strains to back-calculate the force. The weighpads are covered in vulcanized rubber and are attached to a shallow steel foundation frame embedded in a concrete foundation. Bending plate weighpads are available in a number of different sizes, ranging from 20 by 49 inches to 20 by 77 inches.

When used in a WIM system for traffic data collection, two individual weighpads are usually installed in each traffic lane being monitored. The two weighpads are installed in either an inline (side-by-side) or staggered arrangement. All bending plate WIM systems include an inductive loop installed in the pavement some distance before the location of the weighpads in order to detect the presence of a vehicle and initiate the WIM system measurements. A second inductive loop is usually installed in the pavement after the weighpads to detect when a vehicle has moved beyond the weighpads. The staggered installation arrangement permits the vehicle speed to be calculated directly from the two bending plate measurements. The use of the inline arrangement requires an independent axle detection sensor (usually a piezoelectric strip sensor) to compute vehicle speed. The second inductive loop can also be used for this purpose but is generally less accurate for this purpose. The need for a dedicated axle detection sensor for the inline configuration makes this arrangement less desirable from a cost and long-term maintenance perspective than the staggered configuration. Figure 3 shows a bending plate WIM system installation employing a staggered weighpad configuration.

Figure 3. Photo. Staggered bending plate installation and inductive loops. This photo shows a road surface where a bending plate weigh-in-motion system consisting of four weighpads has been installed. Two of the weighpads are staggered left to right, while the other two weighpads are side-by-side down the length of the roadway. A sealant has been used on the road, leaving a tar-like substance along the edges of the weighpads and forming slightly raised outlines.

Figure 3. Photo. Staggered bending plate installation and inductive loops.(33)

Load Cell Sensors

Load cell-based WIM systems use load cells for the weight sensor. A load cell is a transducer that converts an externally applied force into a proportional electrical signal. Although a load cell can be a hydraulic device with a piston and cylinder arrangement, most currently available WIM systems with load cells use strain gage-type sensors. The sensing element of these load cells typically consists of pairs of strain gages mounted to both sides of the web of a specially machined shear beam. When a force is applied to the sensing element, the strain gages measure the principal strains on the beam web, which are used to determine the applied load. The sensing element is capable of measuring high forces, is insensitive to the point of loading, and offers good resistance to side loads.(34)

One type of load cell-based weight sensor used in WIM systems has a single load cell mounted to a steel frame under the center of a rectangular steel loading plate. The weighing unit employs a torque tube that transmits any weight applied to the surface of the load plate to the single load cell. Another type of load cell-based weight sensor used in WIM systems has a total of four load cells installed between a steel frame and each corner of a rectangular steel loading plate. The weight of a wheel located at any position on the loading plate is determined by summing the forces measured by each individual load cell. The rectangular load plates in these systems are approximately 30 by 72 inches, which is large enough to enable each wheel set of a given axle to be weighed individually. Load cell-based systems require a reinforced concrete vault or foundation to support the scales. These vaults are expensive and time-consuming to construct.

As with bending plate systems, load cell-type scales can be installed in a given traffic lane using either an inline or staggered configuration. Inductive loops are usually installed in conjunction with these systems to activate the vehicle measurement sequence. An axle detection sensor may also be used in conjunction with the scales if they are installed using the inline configuration. The load cell-based WIM sensors are the most expensive of the available weight sensor technologies, considering the procurement and initial installation costs; however, they are also considered to be the most accurate WIM technology and are very durable, resilient systems. Figure 4 shows a load cell-based WIM system installed using an inline configuration.

Figure 4. Photo. Inline configuration of load cell-based weight sensors. This photo shows two lanes of a highway. Weighpad sensors have been installed in the road surface. An 18-wheel truck is approaching the camera from the top-right side of the photo in the right lane.

Figure 4. Photo. Inline configuration of load cell-based weight sensors.(33)

Piezoelectric Sensors

Piezoelectric technology can also be used for the weight sensor in a WIM system. Piezoelectric sensors may either be used for measuring weights and vehicle classification or only for vehicle classification purposes (e.g., axle detection and vehicle speed). The sensors used as weight sensors require stricter manufacturing and performance characteristics than those used only for vehicle classification purposes.

Piezoelectric sensors are available in different forms, but they all operate on the principle that when force is applied to a piezoelectric material, a voltage is generated in proportion to the force.(33) The relationship between the applied force and the generated voltage can be quantified and used to determine the weight of a wheel or axle crossing the sensor. This transduction principle only works with dynamically applied loads; weight sensors based on this technology are not suitable for measuring loads applied to them very slowly and cannot be used for static weight measurements. As with bending plate and load cell WIM systems, inductive loops are also used in a WIM application with piezoelectric weight sensors to initiate the measurements when a vehicle’s presence is detected. Many WIM system designs using piezoelectric weight sensors employ two separate lines of the sensors, spaced some distance apart, and installed perpendicular to the lane direction (double threshold system) to increase the accuracy of the weight measurements and to collect vehicle classification data.

There are three basic types of piezoelectric sensors available for WIM applications: piezoceramic sensors, piezopolymer sensors, and piezoquartz sensors.

Piezoceramic Sensors: Piezoceramic sensors consist of ceramic powder compressed between a solid copper core and an outer copper sheath. The sensors have small diameters and are similar in size to regular coaxial cables. When they are used as a weight sensor in a WIM system, they are typically placed in a rigid metal channel filled with a glass fiber-reinforced epoxy resin. The top of the sensor must be installed level with the top of the pavement, which requires it be placed in an approximately 2-inch2 slot cut into the pavement surface and grouted in place. These sensors are available in standard lengths of 6 or 12 ft. Piezoceramic sensors provide weight measurements of average quality and are most suitable for vehicle classification purposes.(33)

Piezopolymer Sensors: Piezopolymer sensors consist of piezoelectric polymer material surrounded by a flat brass casing.(33) The sensor can be installed in a 1-inch-deep by 0.75-inch-wide slot cut into the pavement surface. Installation brackets are placed in the slot at 6-inch intervals to position and support the sensor; once installed, the sensor is surrounded by grout which can be ground flush with the pavement surface after hardening. These sensors have some limitations that reduce the quality of the weight measurements and are more commonly used as axle detectors for vehicle classification purposes.

Piezoquartz Sensors: Piezoquartz sensors are the newest of the piezoelectric sensors available for collecting weight measurements in a WIM system. The piezoelectric material used in this sensor is not sensitive to temperature changes.(33) The design of the sensor packaging is also fundamentally different from the other types of piezoelectric sensors. They are more expensive than the other types of piezoelectric sensors but have been found to be capable of providing good-quality weight measurements. The piezoquartz sensors are installed in an approximately 2-inch-wide slot cut into the surface of the pavement and grouted in place. The sensors are available in different lengths, including 5, 6, and 6.5 ft. Multiple sensors must be installed end-to-end on the same line to cover the full width of a traffic lane. A WIM system design using these weight sensors typically includes a single line of the sensors and two inductive loops, or two separate lines of these sensors and one or two inductive loops. The use of two separate lines of these sensors separated by some distance constitutes a double threshold setup that improves the quality of the weight measurements and collects measurements that can be used for vehicle classification. Piezoquartz sensors were used extensively in the LTPP Specific Pavement Study (SPS) Traffic Data Collection Pooled Fund, TPF-5(004) and found to be capable of providing research-quality traffic load data.(34,36) Figure 5 shows a double threshold installation with these sensors for a WIM system.

Figure 5. Photo. Double threshold WIM system setup with piezoquartz sensors. This photo shows a double threshold weigh-in-motion (WIM system setup with piezoquartz sensors. There is a roadway that has been marked for the installation of two piezoquartz sensors. Using road paint, two large squares have been painted and labelled "L1" and "L2." A separating line has been painted between the two sensor installation sites.

Figure 5. Photo. Double threshold WIM system setup with piezoquartz sensors.(33)

Fiber-Optic Sensors

Fiber-optic sensors have also been explored for use as weight sensors for WIM systems. They can function at high speeds, have low temperature dependency, do not require an electric supply, and have the ability to process data in real time.(36) Although fiber-optic WIM systems have been demonstrated in the field, there are currently no commercially available systems.

Weight Sensor Summary

The different weight sensor technologies described in this report have different advantages and disadvantages with respect to their initial cost, installation requirements, and maintenance needs. To select an appropriate WIM sensor, it is important to consider various criteria, including the following:(38)

Table 2 provides a comparison of weight sensor technologies used with WIM systems for traffic load data collection with respect to their initial cost, expected life, applicability, reliability, and sensitivity of the sensing element to temperature changes from a study published in 2007.(38) The costs for each sensor type have increased since then, but the table provides a valid, relative comparison between relative costs and the advantages and disadvantages of the different systems. In general, bending plates are expensive to procure and install, but they can produce research-quality traffic data. Load cell-based WIM systems are the most expensive to procure and install, but they can provide the most accurate weight data. Conventional piezoelectric sensors provide the lowest accuracy with the lowest relative cost. Quartz-piezoelectric sensors have been shown to provide accurate weight data, but the sensors can be expensive. WIM systems using bending plates and quartz piezoelectric sensors were both shown by the LTPP SPS Traffic Data Collection Pooled Fund Study, TPF-5(004) to be capable of providing high-quality traffic data over a period of years if certain site selection and installation procedures are followed and if periodic maintenance, validation, and calibration of the systems are performed. (34,36)

Table 2. WIM sensors comparison.(38)
Characteristic Bending
Plate
Single
Load Cell
Piezoelectric
Sensor
Quartz-Piezoelectric Sensor
Cost Initial installation cost per lane (USD)
Medium
(~$20,000)
High
(~$50,000)
Low
(~$9,000)
Medium
(~$20,000)
Annual maintenanceand operation
costs(USD)
Medium
(~$6,000)
High
(~$8,000)
Low
(~$5,000)
High
Accuracy (GVW 95-percent confidence)
±10 percent
±6 percent
±15 percent
±10 percent
Sensitivity
Medium
Medium
High
None to temperature, but high to roughness
Expected life (years)
6
12
4
Expected > 15
Reliability
Medium
High
Low
Medium

Vehicle Classification and/or Identification Sensor

Vehicle classification data are needed for pavement and bridge design and rehabilitation as well as for traffic analysis. Most WIM systems also collect data used for vehicle classification. Vehicle classification can be accomplished from the measurements recorded by the weight sensors (if they are configured in a double threshold setup or a staggered configuration) or using a combination of the measurements from the weight sensors (inline or single line of sensors) and a dedicated axle detector also installed in the pavement. The inductive loops used in conjunction with WIM systems to detect vehicle presence and to start and stop the measurement sequence can also be used in order to classify vehicles; however, the loops cannot provide information on vehicle characteristics with the same level resolution that is possible with the other approaches previously described.

Regardless of the WIM system used, identifying vehicle configuration involves careful manual analyses of WIM data, including the use of photographs or video to match the load indications with known or feasible truck axle configurations. Identifying realistic truck configurations is critical for performing certain types of traffic/loading analyses. Typical WIM vehicle classification data validation efforts can successfully identify more than half of the wheel loads recorded.

Inductive Loop Detectors

Inductive loop detectors are typically used to determine the entry and passage of the vehicle from the WIM system. The detectors can also provide various information, including vehicle speed, axle spacing, and vehicle length as the vehicle passes over the WIM station.(39) The primary role of loop detectors in WIM systems is to trigger the system to start and stop the weight measurement and classification sequence for each passing vehicle.

From a technical point of view, an induction loop is an electromagnetic system that uses the relative movements of metal over a wire loop that changes the inductance of the loop. An inductive wire loop is installed in the pavement while an electronics unit transmits energy into its wire loops at certain frequencies, depending on the application. When a vehicle passes over or is stopped within the loop, the vehicle induces currents in the wire loops, decreasing the loop’s inductance and causing the controller connected to the loop to detect the vehicle. A single-loop detector can capture information such as the presence, counts, flow, and lane occupancy of the passing vehicles. Dual-loop detectors (which often use a pair of loops deployed a few feet apart in the same traffic lane) can be used to determine vehicle speed, axle spacing, and average vehicle length. Inductive loops are generally better suited for vehicle detection than they are for providing high-quality vehicle classification information.

Cameras

Some WIM systems incorporate automatic number plate recognition or license plate recognition cameras to read vehicle registration plates to catalog passing vehicles. They can use existing closed-circuit television or road-rule enforcement cameras or ones specifically designed for vehicle surveillance. Such systems commonly use infrared lighting to compensate for headlights and poor weather conditions that might affect recognition any time of day.(40) Pattern recognition technologies for optical characters are applied to images taken by cameras. The processing of these images can be performed entirely at the lane location in real time or later after images for many lanes are transmitted to a remote computer location. The pattern recognition technology tends to be region-specific, owing to plate variation from place to place.

Another type of camera that may be used for WIM systems is the Internet protocol camera, which is a type of digital video camera used to capture a photograph of an entire vehicle from its side and send and receive data via the Internet. Such systems provide high-resolution color images or video that may require a central network video recorder to handle the recording and storage. Photo imaging systems such as these are sometimes integrated with WIM systems used for screening purposes in weight enforcement applications, such as Virtual WIM. Video cameras are not commonly used with WIM systems designed for traffic data collection purposes, although they can be used to evaluate and validate vehicle classification algorithms employed by the WIM system in these instances.

Laser Scanning

Laser scanning technology is used not only for vehicle presence detection, but also for three‑dimensional (3D) vehicle shape recognition and classification (width, length, height, shape,etc.). Such systems, often mounted over the traffic lane, typically emit two eye-safe laser beams to scan the roadway and passing vehicles in order to create a 3D image of the object. Unlike traditional cameras that collect color information about surfaces within its field of view, a 3D scanner collects distance information about surfaces within its field of view. These distances are then used to reconstruct the 3D position of each point on the subject vehicle. When a vehicle enters the beam, the measured distance decreases, and the corresponding vehicle height is calculated based on simple geometry. As the vehicle moves along, the second beam is broken in the same manner. The vehicle speed and length are estimated by measuring the time difference between the breaking of the two beams. Data collected by the sensor are processed and transmitted in real time. This technology can be incorporated with a WIM system design for truck screening in enforcement applications but may be considered too expensive to employ with WIM systems used for traffic data collection.

Automatic Vehicle Classification (AVC)

AVC equipment uses a series of inputs that usually include vehicle presence, number of axles, and the spacing between axles to categorize vehicles into different classes. WIM systems also use these inputs and the axle weights to categorize vehicles. The calibration reexamines the process tests to ensure the algorithm using these inputs correctly classifies the vehicles in WIM stations. Depending on the results, adjustments are made to the algorithm until the output meets the acceptance criteria.

Vehicle weight and classification measurements are made by WIM sensors and AVC systems under challenging, dynamic conditions, and their accuracy and reliability are affected by actual traffic and site conditions at a particular WIM location. No single, standard calibration values exist for these systems that can be valid for all possible locations and traffic conditions. As a result, site-specific calibration and validation procedures must be used for each system. Standardized calibration and validation procedures must be executed following the initial installation of AVC and WIM systems before they can be used to collect reliable traffic data measurements.

AVC calibration involves comparing vehicle classification counts with independent measurements of those same vehicles for a sufficient time period to capture representative data. Independent counts are done manually or by using a video recording technology and then converting the recorded information to classification information. These field tests of the system are the primary means for validating that the AVC equipment and its associated software algorithms are accurately classifying vehicles at a given installation site.

Processor and Data Storage Unit

A processor and data storage unit receives and analyzes the signals from the weight sensors and vehicle classification and/or identification sensors to generate the axle load and vehicle type information that can be used directly by the end user. In most cases, the processor and data storage unit also provides power to the WIM system through a power supply connected to external alternating current power or direct current batteries.(41)

User Communication Unit

The user communication unit provides a communication link between the processor and data storage unit and the user interface. The communication link can be connected directly to a personal computer at the field site or remotely accessed through a wired or wireless modem connection for controlling the systems and transmitting the measurement data offsite. The processor and data storage unit can also directly display the collected data and serve as the user communication unit.(41)

 

 

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