SciELO - Scientific Electronic Library Online

vol.29 issue1Structural damage detection: comparison between GA and PSO techniquesAnalysis of the impact of flooding in the real estate value in lajeado city, brazil: case study for single-family homes author indexsubject indexarticles search
Home Pagealphabetic serial listing  

Services on Demand




Related links

  • On index processCited by Google
  • Have no similar articlesSimilars in SciELO
  • On index processSimilars in Google


Revista ingeniería de construcción

On-line version ISSN 0718-5073

Rev. ing. constr. vol.29 no.1 Santiago  2014 


Design, construction and operation of a prototype to measure load in a flexible pavement structure


Hugo Rondón¹*, Eduardo Delgadillo*, Wilson Vargas*

* Universidad Distrital Francisco José de Caldas, Bogotá. COLOMBIA

Dirección de Correspondencia

This paper presents the design and construction of a prototype able to measure weigh-in-motion on a flexible pavement structure, through the vibrations generated when vehicles move on these road structures. The physical principle is simple: when a vehicle is moving across the pavement road surface, it creates vibrations that can be noticed in areas adjacent to the analyzed highway and the amplitudes of these vibrations can provide information on the magnitude of the vehicle load. Additionally, the paper reports the results of a field study performed in order to calibrate a mathematical equation which simulates the vibration signals obtained by the prototype when vehicles move across the road surface. As a general conclusion, the prototype is able to predict the load that is moving on the pavement, through the vibrations that they induce when transiting on its surface.

Keywords: Spectral load, pavements, vibration, weight in motion – WIM, transit

1. Introduction

Traffic is one of the main variables considered for pavement structures design. This variable is calculated by measuring and weighing vehicles in motion and load axles travelling on a road, during a given period of time (usually one week). Afterwards, these data is calculated by pavement design, so as to determine the number of axles travelling on the road during the complete pavement life span. Each load axle can represent weight, configurations and different tires inflation air pressures, which provide certain degree of aggressiveness and damage (generally measured in terms of fatigue, rutting and erosion). Therefore, in general, for the calculation of the transit variable, this aggressiveness degree of each load axle will theoretically produce a standard axle equivalent to 8.2 t (this axle is derived from results provided by test segments controlled and monitored by AASHO Road Trials, 1961 in Illinois, United States, which are the base of AASHTO methodology, 1986, 1993).

A new methodology for the assessment of pavement performance level uses load spectrums as a traffic variable. In accordance with Castellanos and Rojas (2005), load spectrums are a graphical representation of load per axle travelling on a road, for a given period of time. They are represented by histograms, which are disaggregated per each vehicle type moving across the road. This variable has been employed in recent methodologies for pavement design, for instance: MEPDG (2004) (Prozzi and Hong, 2007, 2007a; Lu et al., 2009; Wang et al., 2011; Haider et al., 2012; Nassiri et al., 2013). This variable could replace the traditional calculation method for equivalent axles of 8.2 t in the near future. Spectral loads consider that loads travelling on a pavement surface have different magnitudes and configurations (some examples may be reviewed on Garnica & Correa, 2004 and Castellanos & Rojas, 2005).

In Colombia, load vehicles weighing used for the calculation of the traffic variable is carried out by using weighing machines. Above procedure has the following disadvantage (Benekohal et al., 2000):

Special weighing facilities are required (generally at toll booths located out of town).
It is difficult to weigh different vehicles at specific points on urban roads.
Weighing is carried out in a static manner. Distribution per axle is considered, but not the dynamical impact of the load axles set.
It is a slow, expensive and dangerous procedure on high traffic highways (Zhang et al., 2008).
Sometimes vehicles exceeding the allowed load limit cannot be identified, because their drivers know the position of load weighing facilities and develop strategies to avoid being noticed. This fact increases vehicles accident probabilities and provokes great damages to road structures (Jacob and de La Beaumelle, 2010).

In order to solve above problems, authors proposed to design and build a prototype capable of measuring the magnitude and the amount of load travelling on a pavement road, by means of vibrations generated by such loads when moving across a tread layer. The physical principle is simple: when a vehicle moves across a pavement surface, it generates vibrations that can be detected by the analyzed road adjacent zones. The amplitudes derived from such vibrations are able to provide information about live-load magnitude. The same principle has been employed by some studies, developed in Bogota D.C., to evaluate bridges deformability by using accelerometers and; in Buenos Aires (Kantor & Pérsico, 2005) to evaluate the effect produced by vibration generated by vehicular loads on buildings structural behavior.

Results reported by the present study correspond to the first stage of a research intended to design and build a prototype tool capable of calculating the loads moving across a flexible pavement road without stopping vehicular traffic. Load amplitudes generated by vehicles on the road, are calculated by using vibration amplitudes they produce when travelling on a pavement surface. Future stages of this research will focus in improving the prototype, so that it will be able to measure, in a more reliable way, the loads traveling on any flexible pavement structure exposed to different factors that might affect measurements (environmental conditions, pavement condition, load types, etc.). By designing and building the prototype, authors intend to develop a new technology to calculate vehicular load moving across flexible pavement structures in Colombia. Additionally, the prototype will help to:

1.       Detect illegal circulation of trucks exceeding maximum loads allowed by the Transportation Ministry in Colombia.
2.       Collect information that might be used by organizations that design and manage roads network in Colombia.

This paper initially presents a summary of the most popular devices employed worldwide to estimate pavement loads. Then, authors briefly describe the developed prototype and the results obtained from its on-site utilization on a flexible pavement structure (this kind of structure was chosen because it is the most available in Colombia). Finally, a mathematical equation is introduced, which calculates mass or load moving across the pavement, based on vibrations generated by vehicles in motion.

2. Measurement Systems available Worldwide

Diverse systems have been implemented worldwide in order to calculate loads travelling on pavement structures. The most popular systems are the ones calculating the amount and magnitude of vehicular load without stopping traffic circulation. This kind of technology is known as "weigh in motion" (WIN). The present state of the art knowledge in this area is that, although WIM technology has been studied since the 50´s, it has not been possible to determine a definitive device configuration to accurately predict the magnitude and distribution of live-loads on a road (COST 323, 1999; Liu et al., 2006, Prozzi et al., 2008; Haider et al., 2012). In spite of its great advantages for the determination of vehicular loads moving across road structures, WIM has some disadvantages: current available techniques are not easy to install and to maintain and they are not accurate enough (Prozzi and Hong, 2007, 2007a; Zhang et al., 2008). Furthermore, in some occasions, this technology underestimates loads imposed by the fleet of vehicles, thus leading to an overestimation of road structures life-span (Haider et al., 2012).

There are several types of WIM sensors. The most employed by roads studies are: bending plate sensor, load cells sensor, piezoelectric sensor, capacitance sensors and optical fiber sensors (Teral, 1998; Nikolaidis, 2002; Wang, 2005; Yannis and Antoniou, 2005; Cheng et al., 2007).
A bending plate is a steel-made plate with strain gauges installed underneath the plate (Figure 1). When a vehicle is driven on the road surface, the strain induced by the vehicular load can be measured and turned into a dynamical load. This kind of sensor can be used on roads where vehicles run at high and low speed. The accuracy of measurements taken by this sensor on the road site is high. Disadvantages: the sensor is not easy to install nor easy to maintain and it is an expensive device. Additionally, when measures are carried out on flexible pavement structures, it is necessary to build a concrete foundation to support the bending plate sensor (see Figure 1).

Figure 1. Bending plate (Florida Department of Transportation, 2007)

According to Liu et al. (2006), load cells sensors are the most sensitive and accurate devices to measure dynamical loads, among WIM sensors. These kinds of sensors are assembled in the center under a framework with an armored steel camera located on the road grade beam. Cells are assembled in such a way that vehicle wheel rims will circulate on them. This is an expensive and difficult to install device. The cell that measures load and the circuit where it is installed is schematically shown by Figure 2.

Figure 2. Load cells scheme. Liu et al. (2006)

The piezoelectric sensor (Figure 3) is generally employed to measure loads at high speed. At low speed or low static loads, measurements carried out by this device are quite unsatisfactory. The main advantages of this sensor are the low operational costs, in comparison to other devices, and it is easy to install and easy to work with. The main disadvantage of using piezoelectric sensors is that their measurements are usually not accurate enough. This sensor is made of a material that generates static charge when it is mechanically strained. Its name comes from the Greek word piezo that means "I apply pressure". This type of sensor can be installed inside the pavement for use on permanent basis or, installed on the road to be used as a portable device.

Figure 3. Piezoelectric sensor. Liu et al. (2006)

Normally, piezoelectric devices are composed by polymer molecular chains, such as polyvinylidene fluoride, ceramics or crystals like quartz. Piezoelectric sensors are usually coaxial with a metal base and piezoelectric material covered by an external metal layer (Martin et al., 2003).

A capacitance sensor (Figure 4) has two or more metallic plates installed in parallel, electric conductors at different loads are installed between them. If a vehicle presses this sensor, the distance existing between plates decreases but capacitance property increases. Given such information on distance and capacitance, the calculation of axle load circulating on the road is obtained. Generally, the materials used to elaborate such elements are stainless steel, aluminum, polyurethane and rubber.

 Figure 4. Capacitance sensor. Liu et al. (2006)

The main advantage of optical fiber sensors is that they can be employed where other WIM sensors cannot (for example bridges, near train ways or near power stations) (Yannis and Antoniou, 2005). Besides optical fiber yields better results and accuracy when calculating live- loads, in comparison to piezoelectric devices (Federal Highway Administration, 2001). A scheme of optical fiber sensor is shown by Figure 5.

Figure 5. Optical fiber sensor. Gan et al. (2008)

The state of the art knowledge and more detailed bibliography about this subject may be reviewed on ASTM (1994), Cropley (1996), Bushman and Pratt (1998), Whitford (1998), Peters and Koniditsiotis (2000), Beckett et al. (2010).

3. Description of prototype

The block diagram on Figure 6 shows the stages developed to build the prototype. In this figure, five main stages are shown: sensor, signal conditioning, a/d converter; PSoC processor (which carries out data collection, adaptation, and it processes the sensor signal to measure load spectrum) and PC.

Figure 6. Blocks diagram of monitoring system

The sensor employed is an accelerometer of type mems MMA7260QT free-scale semiconductor of three axles, which provide an analog signal with a sensitivity of 66mv per gravity. This sensor is capable of measuring accelerations from 1.0g up to 6g. It operates within a voltage range varying from 2.4v to 3.6v. The accelerometer design is capable to obtain an electric signal, which is proportional to the acceleration on the surface it has been installed.

The programmable systems-on-chip (PSoC) are a configurable "mixed signal array" (partly analogical and digital), controlled by a board. Such devices combine the advantages of SoC, with the flexibility of programmable systems. These microcontrollers have been designed to replace multiple and traditional MCU by a single programmable device, at low cost. The PSoC include analogical and digital programmable blocks configured by software. This architecture allows the user to create optimized peripheral configurations
to meet the requirements of each "Embedded System".

Besides, a powerful CPU, programmable flash memory, SRAM data memory and programmable E/S ports are included.

For the development of this prototype, it was necessary to standardize the signal obtained by the accelerometer, which captures the vibrations produced on the pavement surface when vehicles move across the tread-layer. Accelerator output signals were measured, thus obtaining signals as the one shown by Figure 7.

 Figure 7. Sensor signal

Signal amplitude obtained by the accelerometer is wirelessly transmitted into a PC. The signal is then displayed by the PC screen by using the LabView data collection software. The acquired signal enters the PSoC microcontroller and is amplified, and then it turns from analogical into digital by using an A/D converter built in the PSoC. The signal is filtered and then transmitted by the data communication module in the PSoC. Figure 8 shows a flow diagram for the signal treatment in the PSoC.

Figure 8. Flow Diagram for the signal treatment in the PSoc

Figure 9 shows the prototype and its components.

Figure 9. Prototype Device developed to detect vibrations

4. Metodology

The project was developed in four phases.

Phase 1. Bibliographical review. This stage compiled and analyzed information about available procedures and technology to measure load spectrums on pavement structures.

Phase 2. Prototype design and construction. This stage corresponds to the prototype design and construction process to measure load spectrums on site. Intrinsic variables were taken into account during this phase (circulation speed, distance between the vehicle and the prototype, vehicle mass, etc.). Extrinsic variables were also considered (tread-layer noise, rain and wind taking place during measurements, etc.), which might affect vibrations measurements.

Phase 3. Experimental phase. Four heavy types of vehicles were weighed (Renault Logan car, middle size bus (usually called buseta) and trucks C2 and C3 type, see Figure 10) so as to accurately obtain the right amount of loads, before tests were applied. A truck weighing machine installed was in static position, which is property of an asphalt company. The measured vehicles loads are 1113, 12334, 16355 and 26015 kg for the car, the middle size bus and trucks type C2 and C3, respectively.

The car distributes loads to the pavement from two simple axles, composed of two wheels each (one on each corner). The bus and C2 truck distribute loads from two simple axles; one forward axle composed by two wheels (one on each corner) and the rear axle composed by four wheels (two on each corner). C3 truck distributes load from a simple forward axle composed by two wheels (one on each corner) and a tandem rear axle composed by two articulated axles each, with four wheels (two on each corner). These vehicles were selected because they are the most driven type of vehicles moving across Colombian road network. Besides, authors wanted to evidence the prototype sensitivity when measuring vehicles vibrations produced by medium size loads, in this first research stage.

Figure 10. Vehicles employed by this study

During the research, vehicles moved across a newly constructed pavement structure, which is composed of a 12 cm asphaltic layer, supported on 20 cm base, 30 cm granular sub-base, 30 cm gravel-surfaced, and CL clay type sub-grade according to the Standardized System for Soil Classification. This structure was selected because the surface was in perfect conditions regarding uniformity and serviceability. Additionally, the structure is confined by a 60 cm width-roadside; the prototype was installed on its center line. The average environment temperature was 20°C on sunny days, during experimental tests.

In order to assess distance influence between the prototype and the analyzed axle load, vehicles were made to move across at approximately 30, 60, 100 and 150 cm horizontal distance (D) between the closest vehicle wheel rim and the prototype. Vehicles were made to move across at such distances and at 30, 60 y 100 km/h in order to assess the speed influence (V). A metallic protection element was installed so as to avoid that wind produced by vehicles at high speed could affect measurements.

Five measurements were carried out for each distance/speed. A benchmark was used to compare loads moving across the pavement structure, which is the average for the maximum vibration amplitude of gravity force applied by each axle when travelling in front of the prototype. That is to say, each vehicle produced two vibration amplitude magnitudes (which were averaged), which were recorded by the sensor when vehicles front axle and rear axle circulated on the pavement structure. The experimental phase was carried out on a flexible pavement structure.

Phase 4. Mathematical model. The results obtained from phase 3, were used to develop an empirical regression equation to calculate the load moving across a pavement structure by employing vibration measured by the prototype. This equation was defined among different equations by using mathematical programs TCWin and techdig.

5. Results

Figures 11-14 show the results obtained from the execution of phase 3, which is described by the Methodology chapter. There is an expected increment of measured amplitude values as long as the vehicle increases load and circulation speed. Additionally, we observe that such amplitude is higher as long as the vehicle is closer to the prototype.

 Figure 11. Evolution of Vvibration amplitude due to speed D=30cm

Figure 12. Evolution of vibration amplitude due to speed D=60cm

Figure 13. Evolution of vibration amplitude due to speed D=100cm

Figure 14. Evolution of vibration amplitude due to speed D=150cm

The results for C3 trucks are not presented by Figure 11, because the amplitude magnitude is higher than 6g, and the sensor range oscillates between 0.1g and 0.6g only. Similarly, Figure 14 does not present the vibration amplitude results obtained for the car, since its vibration amplitude was too low (lower than 0.1g). Above indicates that the prototype is capable of measuring vibration amplitudes from vehicles travelling on the flexible pavement structure at speeds between 30 and 100 km/n and approximated mass between 10 and 17 tons, at distances between 30 cm and 1.5 m.

The best achieved vibration amplitude with the best repetition degree and the least deviation degree was the one obtained when vehicles circulated with a mass of 10 and 17 tons, at 30 cm distance from the prototype and at speed of 100 km/h (closer distance and higher speed). At 60, 100 and 150 cm distance from the prototype, an average decrease is reported for the vibration amplitude of 37, 56 and 74% in comparison to the one obtained at 30 cm. Similarly, such amplitude decreased in an average of 37% when circulation speed decreased from 100 km/h down to 30 km/h.

Figures 15-18 show an improved trend of abovementioned results, when the vibration amplitude is normalized due to the vehicle mass. Such results may be mathematically described by means of Equation 1.


A is vibration amplitude in force gravity (g); m is the vehicle mass in kg; D is the horizontal distance in cm between the vehicle wheel rim, which is closest to the prototype and; V is the vehicle circulation speed expressed in km/h. Figures 15-18 present a simulation of results obtained by Equation (1). There is a good correlation (coefficient r2 between 0.84 and 0.92) between achieved results and those simulated by the Equation 1.


Figure 15. Evolution of the relationship of vibration amplitude with vehicle mass due to speed D=30cm

Figure 16. Evolution of the relationship of vibration amplitude with vehicle mass due to speed D=60cm

Figure 17. Evolution of the relationship of vibration amplitude with vehicle mass due to speed D=100cm


Figure 18. Evolution of the relationship of vibration amplitude with vehicle mass due to speed D=150cm

6. Conclusions

In the present study, an electronic device was designed and built to measure vibrations generated by different vehicular loads moving across a flexible pavement structure. Based on an experimental phase executed on-site, we can conclude that vibrations measured by the prototype can be used to calculate vehicular load moving across the pavement structure when load moves at speeds between 30 and 100 km/h, with an approximated mass between 10 and 17 tons, at distances between 30 cm and 1.5m.

At the same time, a mathematical regression equation was developed. This equation is capable of reliably calculating the mass moving across the analyzed pavement structure (correlation coefficient r² between 0.84 and 0.92), by using the vibration amplitude captured by the sensor, as well as the vehicle-in motion speed and the distance between the vehicle and the prototype. The main disadvantage of this prototype is that it is unable of calculating static charges when the circulation speed is lower than 30 km/h, because at this speed vehicles do not generate vibration that can be captured by the sensor.

In a further study phase, which is currently beginning; the prototype shall be improved so as to measure loads-in-motion lower than 10 tons (cars and buses) and higher than 17 tons (truck tractors, 4-axle trucks or higher). Additionally, the influence from the following factors shall be evaluated (which were not included in this study) about vibrations measured by the prototype: pavement serviceability conditions (type of pavement, surface roughness, surface damages, deflections, material properties, sub-grade type, side confinement, etc.); those ones related to the vehicle (speed, acceleration, kind of wheel rim, load magnitude, kind of suspension, etc); and finally those ones related to environmental conditions (rain, temperature, humidity, wind, etc).


7. References


AASHO - American Association of State Highway Officials. (1961), The AASHO Road Test, Special Report 61 (7 reports, A-G), National Academy of Sciences - National Research Council, Washington D.C.         [ Links ]

AASHTO – American Association of State Highway and Transportation Officials. (1986, 1993), Guide for Design of Pavement Structures, Washington, D.C.         [ Links ]

ASTM - American Society for Testing and Materials (1994), Standard Specification for Highway Weigh-In-Motion (WIM) Systems with User Requirements and Test Method, ASTM E 1318 - 94.         [ Links ]

Beckett N., Bañez B., Darwin D. (2010), Annual Weigh-In-Motion (WiM) Report 2010, NZ Transport Agency (NZTA) Report.         [ Links ]

Benekohal R. F., El-Zohairy Y. M., Wang S. (2000), Truck Travel Time Around Weigh Stations: Effects of Weigh in Motion and Automatic Vehicle Identification Systems. Transportation Research Record, 1716, 135-143.         [ Links ]

Bushman R., Pratt A. J. (1998), Weigh In Motion Technology - Economics and Performance. Presented at the North American Travel Monitoring Exhibition and Conference, Charlotte NC May 11-15, 1998 NATMEC '98.         [ Links ]

Castellanos A. P., Rojas J. P. (2005), Espectros de Carga Vehicular en la Zona Urbana de Bogotá D.C., M. Sc. Thesis, Departamento de Ingeniería Civil y Ambiental, Universidad de Los Andes (Colombia).         [ Links ]

Cheng L., Zhang H., Li Q. (2007), Design of a Capacitive Flexible Weighing Sensor for Vehicle WIM System, Sensors, 7, 1530-1544.         [ Links ]

COST 323 (1999), European WIM Specification, August 1999.         [ Links ]

Cropley S. (1996), Weigh-In-Motion (WIM) Systems: Use in Pavement Design, In: Proc. Weigh-in-Motion Symposium, May 1996, ARRB Transport Research Ltd.         [ Links ]

Florida Department of Transportation, Transportation Statistics Office. (2007), Traffic Monitoring Handbook.         [ Links ]

Gan J.-L., Cai H.-W., Geng J.-X., Pan Z.-Q., Qu R.-H., Fang Z.-J. (2008), Optic Fiber-Based Dynamic Pressure Sensor, Journal of Electronic Science and Technology of China, Vol. 6, No. 4, 482-485.         [ Links ]

Garnica P., Correa A. (2004), Conceptos Mecanicistas en Pavimentos, Publicación técnica No. 258, Secretaría de Comunicaciones y Transportes, Instituto Mexicano del Transporte, Sanfandila, Qro.         [ Links ]

Haider S. W., Harichandran R. S., Dwaikat M. B. (2012), Impact of Systematic Axle Load Measurement Error on Pavement Design Using Mechanistic-Empirical Pavement Design Guide, Journal of Transportation Engineering, Vol. 138, No. 3, 331-386.         [ Links ]

Jacob B., de La Beaumelle V. F. (2010), Improving Truck Safety: Potential of Weigh-In-Motion Technology, IATSS, Research 34, 9–15.         [ Links ]

Kantor P., Pérsico D. (2005), Estudio de Vibraciones en Edificios, Ingeniería Estructural, Buenos Aires, No. 33.         [ Links ]

Liu R., Chen X., Li J., Guo L., Yu J. (2006), Evaluating Innovative Sensors and Techniques for Measuring Traffic Loads. Report 0-4509-1, Project Number: 0-4509, Subsurface Sensing Laboratory, Department of Electrical and Computer Engineering, University of Houston.         [ Links ]

Lu Q., Zhang Y., Harvey J. (2009), Growth of Truck Traffic Volume for Mechanistic-Empirical Pavement Design, International Journal of Pavement Engineering, Vol. 10, No. 3, 161-172.         [ Links ]

Martin P. T., Feng Y., Wang X. (2003), Detector Technology Evaluation, MPC-03-154, the Mountain-Plains Consortium (MPC), November 2003.         [ Links ]

MEPDG. ARA, Inc., ERES Consultants Division (2004), Guide for the Mechanistic-Empirical Design of New and Rehabilitated Pavement Structures, NCHRP Project 1-37A, Transportation Research Board, Washington, D.C., Accessed September 12 of 2007.         [ Links ]

Nassiria S., Bayata A., Kilburn P. (2013), Traffic Inputs for Mechanistic-Empirical Pavement Design Guide Using Weigh-In-Motion Systems in Alberta, International Journal of Pavement Engineering (Accepted to Publish in 2013).         [ Links ]

Nikolaidis D. R. (2002), Real-Time Speed, Classification, and Weigh-In-Motion Using a Single, Spatially Distributed Fiber-Optic Sensor, A dissertation of Florida Institute of Technology.         [ Links ]

Federal Highway Administration. (2001), Truck Weight Monitoring. May 1, 2001,         [ Links ]

Peters B., Koniditsiotis Ch. (2000), Weigh-In-Motion Technology, Austroads Project No. RUM.TM.9, Austroads Publication No. AP–R168/00.         [ Links ]

Prozzi J. A., Hong F. (2007), Effect of Weigh-in-Motion System Measurement Errors on Load-Pavement Impact Estimation, Journal of Transportation Engineering, Vol. 133, No. 1, 1-10.         [ Links ]

Prozzi J. A., Hong F. (2007a), Optimum Statistical Characterization of Axle Load Spectra Based on Load-Associated Pavement Damage, International Journal of Pavement Engineering, Vol. 8, No. 4, 323-330.         [ Links ]

Prozzi J., Hong F., Leung A. (2008), Effect of Traffic Load Measurement Bias on Pavement Life Prediction: A Mechanistic-Empirical Perspective, Transportation Research Record: J. of the Transportation Research Board, 2087, 91-98.         [ Links ]

Teral S. R. (1998), Fiber Optic Weigh-In-Motion: Looking Back and Ahead, Proc. Optical Engineering, Vol. 3326, 129-137.         [ Links ]

Wang K. (2005), A Fiber-Optic Weigh-In-Motion System Based on Fiber Bragg Grating Technologies, A Dissertation of Stevens Institute of Technology, Hoboken, New Jersey, USA.         [ Links ]

Wang K. C. P., Li Q., Hall K. D., Nguyen V., Xiao D. X. (2011), Development of Truck Loading Groups for the Mechanistic-Empirical Pavement Design Guide, Journal of Transportation Engineering, Vol. 137, No. 12, 855-862.         [ Links ]

Whitford R. K. (1998), Truck Weight Monitoring Plan Using Weigh-In-Motion Devices: Plan for WIM for the State of Alaska, Presented at the North American Travel Monitoring Exhibition and Conference, Charlotte NC May 11-15, 1998 NATMEC '98.         [ Links ]

Yannis G., Antoniou C. (2005), Integration of Weigh-in-Motion Technologies in Road Infrastructure Management, ITE Journal, Vol. 75, No. 1, 39-43.         [ Links ]

Zhang W., Suo Ch., Wang Q. (2008), A Novel Sensor System for Measuring Wheel Loads of Vehicles on Highways, Sensors, 8, 7671-7689.         [ Links ]

Fecha de Recepción: 20/03/2013 Fecha de Aceptación: 01/11/2013