National Academies Press: OpenBook

Protocols for Network-Level Macrotexture Measurement (2021)

Chapter: Chapter 2 - Macrotexture Measurement Technologies and Parameters

« Previous: Chapter 1 - Introduction
Page 24
Suggested Citation:"Chapter 2 - Macrotexture Measurement Technologies and Parameters." National Academies of Sciences, Engineering, and Medicine. 2021. Protocols for Network-Level Macrotexture Measurement. Washington, DC: The National Academies Press. doi: 10.17226/26225.
×
Page 24
Page 25
Suggested Citation:"Chapter 2 - Macrotexture Measurement Technologies and Parameters." National Academies of Sciences, Engineering, and Medicine. 2021. Protocols for Network-Level Macrotexture Measurement. Washington, DC: The National Academies Press. doi: 10.17226/26225.
×
Page 25
Page 26
Suggested Citation:"Chapter 2 - Macrotexture Measurement Technologies and Parameters." National Academies of Sciences, Engineering, and Medicine. 2021. Protocols for Network-Level Macrotexture Measurement. Washington, DC: The National Academies Press. doi: 10.17226/26225.
×
Page 26
Page 27
Suggested Citation:"Chapter 2 - Macrotexture Measurement Technologies and Parameters." National Academies of Sciences, Engineering, and Medicine. 2021. Protocols for Network-Level Macrotexture Measurement. Washington, DC: The National Academies Press. doi: 10.17226/26225.
×
Page 27
Page 28
Suggested Citation:"Chapter 2 - Macrotexture Measurement Technologies and Parameters." National Academies of Sciences, Engineering, and Medicine. 2021. Protocols for Network-Level Macrotexture Measurement. Washington, DC: The National Academies Press. doi: 10.17226/26225.
×
Page 28
Page 29
Suggested Citation:"Chapter 2 - Macrotexture Measurement Technologies and Parameters." National Academies of Sciences, Engineering, and Medicine. 2021. Protocols for Network-Level Macrotexture Measurement. Washington, DC: The National Academies Press. doi: 10.17226/26225.
×
Page 29
Page 30
Suggested Citation:"Chapter 2 - Macrotexture Measurement Technologies and Parameters." National Academies of Sciences, Engineering, and Medicine. 2021. Protocols for Network-Level Macrotexture Measurement. Washington, DC: The National Academies Press. doi: 10.17226/26225.
×
Page 30
Page 31
Suggested Citation:"Chapter 2 - Macrotexture Measurement Technologies and Parameters." National Academies of Sciences, Engineering, and Medicine. 2021. Protocols for Network-Level Macrotexture Measurement. Washington, DC: The National Academies Press. doi: 10.17226/26225.
×
Page 31
Page 32
Suggested Citation:"Chapter 2 - Macrotexture Measurement Technologies and Parameters." National Academies of Sciences, Engineering, and Medicine. 2021. Protocols for Network-Level Macrotexture Measurement. Washington, DC: The National Academies Press. doi: 10.17226/26225.
×
Page 32
Page 33
Suggested Citation:"Chapter 2 - Macrotexture Measurement Technologies and Parameters." National Academies of Sciences, Engineering, and Medicine. 2021. Protocols for Network-Level Macrotexture Measurement. Washington, DC: The National Academies Press. doi: 10.17226/26225.
×
Page 33
Page 34
Suggested Citation:"Chapter 2 - Macrotexture Measurement Technologies and Parameters." National Academies of Sciences, Engineering, and Medicine. 2021. Protocols for Network-Level Macrotexture Measurement. Washington, DC: The National Academies Press. doi: 10.17226/26225.
×
Page 34
Page 35
Suggested Citation:"Chapter 2 - Macrotexture Measurement Technologies and Parameters." National Academies of Sciences, Engineering, and Medicine. 2021. Protocols for Network-Level Macrotexture Measurement. Washington, DC: The National Academies Press. doi: 10.17226/26225.
×
Page 35
Page 36
Suggested Citation:"Chapter 2 - Macrotexture Measurement Technologies and Parameters." National Academies of Sciences, Engineering, and Medicine. 2021. Protocols for Network-Level Macrotexture Measurement. Washington, DC: The National Academies Press. doi: 10.17226/26225.
×
Page 36
Page 37
Suggested Citation:"Chapter 2 - Macrotexture Measurement Technologies and Parameters." National Academies of Sciences, Engineering, and Medicine. 2021. Protocols for Network-Level Macrotexture Measurement. Washington, DC: The National Academies Press. doi: 10.17226/26225.
×
Page 37
Page 38
Suggested Citation:"Chapter 2 - Macrotexture Measurement Technologies and Parameters." National Academies of Sciences, Engineering, and Medicine. 2021. Protocols for Network-Level Macrotexture Measurement. Washington, DC: The National Academies Press. doi: 10.17226/26225.
×
Page 38
Page 39
Suggested Citation:"Chapter 2 - Macrotexture Measurement Technologies and Parameters." National Academies of Sciences, Engineering, and Medicine. 2021. Protocols for Network-Level Macrotexture Measurement. Washington, DC: The National Academies Press. doi: 10.17226/26225.
×
Page 39
Page 40
Suggested Citation:"Chapter 2 - Macrotexture Measurement Technologies and Parameters." National Academies of Sciences, Engineering, and Medicine. 2021. Protocols for Network-Level Macrotexture Measurement. Washington, DC: The National Academies Press. doi: 10.17226/26225.
×
Page 40
Page 41
Suggested Citation:"Chapter 2 - Macrotexture Measurement Technologies and Parameters." National Academies of Sciences, Engineering, and Medicine. 2021. Protocols for Network-Level Macrotexture Measurement. Washington, DC: The National Academies Press. doi: 10.17226/26225.
×
Page 41

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

24 Based on the literature review and survey of state highway agencies, the research team iden- tified, summarized, and analyzed current and emerging measurement technologies for collecting macrotexture data at the network level. The team also identified and evaluated available param- eters for characterizing macrotexture. The results of the literature review and survey were complemented by contacting equipment manufacturers and service providers to further assess the various technologies, methods, and approaches. 2.1 Available and Emerging Macrotexture Measurement Technologies There are several technologies currently available to collect macrotexture data. Some of the available equipment, such as the single-spot laser system, has been used for years to gather macrotexture data and has a well-established track record. Other types of equipment, such as line lasers and 3D laser and optic systems, have been in development and are just beginning to make their debuts in network-level data acquisition. All of these technologies can generally be categorized into three groups: stationary, walking speed, and highway speed. 2.1.1 Stationary These devices require the data collection team to hand-place the equipment. A portable device is placed on the pavement surface, and the measurement is made. Several measurements are required on homogenous pavement surfaces to find the average characteristics of the surface. Operator judgment is needed to determine how many samples are collected, where the samples are collected, and what constitutes a “homogeneous pavement section.” These types of devices always require traffic control and pose a significant risk from oncoming traffic to those collecting the data. These devices, however, are generally regarded as ground truth devices. They can measure the surface macrotexture with a high degree of accuracy because they use a physical medium (such as sand) to discretely measure macrotexture, can record several samples at different orientations in one placement (e.g., C.T. Meter), or sample a near-continuous 3D profile of the pavement. Stationary devices generally fall into three categories: volumetric, 2D profiles, and 3D profiles. Volumetric devices use a known volume of a physical medium, such as sand, glass spheres, or grease, to fill the pavement macrotexture. The pavement area required to contain the physical medium is computed, and an average texture depth is calculated by dividing the volume of the medium by the surface area covered by the medium. The smaller the area covered by the physical medium for the same volume of physical medium, the greater the macrotexture depth of the surface. A variation of the volumetric procedure is the outflow meter, which uses a known volume of water in a graduated cylinder. Unlike other volumetric methods where an area is C H A P T E R   2 Macrotexture Measurement Technologies and Parameters

Macrotexture Measurement Technologies and Parameters 25   computed, the time required to completely drain the water from the cylinder is recorded and this value can be correlated to other macrotexture parameters. Outflow devices have the benefit of quantifying the benefit of subsurface interconnected voids from surfaces such as porous or pervious pavements and OGFC. Two-dimensional profiles were first recorded using a stylus that was moved across the pave- ment surface, recording a continuous profile of the amplitude of the surface macrotexture. Today, a laser is typically employed to make these same measurements, eliminating errors due to mechanical limitations of stylus devices. Two-dimensional profiles have become the de facto standard used in gathering macrotexture data, given the most prevalent technologies used to collect macrotexture data are 2D (i.e., single-spot lasers). This has given rise to the prevalence of MPD as the parameter of choice for describing a pavement’s macrotexture. Examples of equip- ment that collect stationary 2D profiles include the commonly used C.T. Meter. Three-dimensional surface maps can also be sampled and recorded to measure pavement macrotexture. The most common methods to collect this information are laser texture scanners, stereoscopic vision, or some combination of the two. A clear advantage of 3D systems over 2D stationary profiles is the larger quantity of data gathered over a larger surface area. This additional data can provide more information about a pavement surface, such as directionality, spacing, and orientation of macrotexture asperities, and can be used for computation of macrotexture parameters, such as the MTD. Three-dimensional profiles, however, also take longer to capture, can require more data post-processing to interpret the results of the scan, and are less widely used at present. 2.1.2 Walking Speed These devices often use the same types of sensors as stationary devices; however, an operator pushes or guides the device at walking speed. An advantage over stationary devices is that a larger sample of the pavement macrotexture is gathered as the operator walks along the pavement surface with the device. Some devices can only collect 2D data, while others can collect 3D data. These devices are typically used to collect project-level macrotexture data, data on a section of interest (i.e., in response to an accident), data for calibration of other data collection devices, or can be used anywhere small-scale data collection is needed, such as in research. 2.1.3 Highway Speed These devices collect near-continuous pavement surface profiles at high speed. The sensors used to collect the data are mounted to a vehicle, and data on the road network can be collected at highway speeds in a single run. These devices represent the most feasible solution for the network-level data collection. No traffic control is required, given the vehicles travel at the speed of traffic and use contactless sensors. At the time of conducting the research, no such device had proved to have the required resolution and sampling rate combination to gather the 3D macrotexture data useful for the computation of existing macrotexture parameters. Therefore, the research focused on the current practice of collecting 2D data in the longitudinal (traffic) direction using single-spot or line lasers. Table 4 was developed based on an extensive review of available literature, survey results from state departments of transportation (DOTs), interviews with sensor manufacturers, and macrotexture measurement service providers. The table summarizes current and emerging technologies for measuring pavement macrotexture at the network level. To be considered for inclusion in this list, the technology had to either be currently used to collect macrotexture data or at near readiness for commercial application.

Commonly Used Single-Point Lasers Product Name Manufacturer Technology Laser Wavelength (nm) Maximum Sampling Rate (kHz) Standoff (mm) * Measurement Range (MR) (mm) Vertical Resolution (% MR) Maximum Vertical Resolution (mm) Average Laser Spot Size (mm) Laser Power (mW) Gocator 13xx LMI Technologies Single-Spot Laser 660 32 200–562 200–375 0.00075– 0.001% 0.00150 0.5–1.8 3B (< 130) Optocator 20xx LMITechnologies Single-Spot Laser 780 78 180–1200 32–1024 0.003% 0.00096 0.2 20 RoadRun System LIMAB Single-Spot Laser 635–670 62 100–300 200–1000 0.001– 0.005% 0.00200 1.0 20–40 AR 700 Series Acuity(Ames partner) Single-Spot Laser 670 100 12.7–1422 3.175–1270 0.005% 0.00020 0.1 1–20 Emerging Technologies Product Name Manufacturer Technology Laser Wavelength (nm) Maximum Sampling Rate (kHz) Standoff (mm) * MR (mm) Vertical Resolution (% MR) Maximum Vertical Resolution (mm) Field of View (mm) Data Points per Profile Gocator 2342 LMITechnologies Line Laser and Sensor 660 5 190 210 0.0079– 0.0190% 0.01659 64–140 1280 Gocator 2375 LMITechnologies Line Laser and Sensor 808 (near infrared) 5 650 1350 0.0114– 0.0415% 0.15390 345–1028 1200 LCMS Pavemetrics Line Laser with Line Scan Camera Invisible 11.2 Near Roof 250 0.10000% 0.25000 2000 2080 LCMS II Pavemetrics Line Laser with Line Scan Camera Invisible 28 Near Roof 250 0.04000% 0.10000 2000 2080 AR 500 Series Acuity Single-Spot Laser Distance Triangulation 405 (blue) 100 12.7–1422 3.175–1270 0.005% 0.00020 0.1 1–20 * The standoff distance in a laser system is the distance from the laser emitter to the surface that it is intended to measure (pavement surface). Table 4. Current and emerging measurement technologies.

Macrotexture Measurement Technologies and Parameters 27   The most common devices are HSLE systems that use the single-spot laser triangulation method. Under this approach, a single laser beam (generally smaller than 1 mm in diameter at its standoff distance in the center of its measurement range) is emitted from the device. The light is reflected from a discrete point of the pavement to the device’s light sensor (typically a charged coupled device or complementary metal-oxide semiconductor [CMOS] sensor like the one in a digital camera). Triangulation is then used to determine the distance between the instrument and the pavement surface at the point measured. Distance triangulation can be per- formed on any right triangle if at least one angle (based on the angle of the sensor relative to the laser) and one distance (the orthogonal distance from the laser beam to the position on the sensor) is known. Line lasers operate on a similar principle; however, the laser is spread into a beam through optics before striking its target, and a 2D sensor is used to record many individual center-weighted light points. Triangulation is then carried out for these many distances and angles (see Figure 12). Profiles from single-spot lasers can be defined as “continuous” if the sampling rate of the sensor is high enough to provide continuous coverage (based on the laser spot size) of the surface at the speed driven. Sampling rates as high as 100,000 samples per second (100 kHz) have been reported by various manufacturers. However, faster sampling rates often result in higher signal noise (Schleppi et al. 2016), especially when used on challenging surfaces such as dark and/or reflective pavements. If the sampling rate is too slow (i.e., the device is not “fast” enough), information about the shape of pavement asperities can be lost, as demonstrated by Liu et al. (2016). The measurement method on the horizon for macrotexture is 3D. The general approach for collecting 3D macrotexture data is by use of a line laser and optical sensor. Under this scheme, laser light is emitted from a diode (just as with the single-spot laser method); however, the beam is diffused into a line that is projected on the pavement’s surface and subsequently reflected to the device’s sensor. On the sensor, the reflected laser light is handled much in the same way as typical laser triangulation, but en masse. Instead of a single point on the sensor, the line covers many pixels of the sensor, and each is taken as an individual data point. The number of data points taken per profile is dependent on the sensor used (how many pixels per row). The laser line’s width (field of view) is determined by the optical diffuser used and the mounting height of the device. Figure 12. Diagram of single-spot laser triangulation method.

28 Protocols for Network-Level Macrotexture Measurement Table 5 lists companies that employ the sensors listed in Table 4 to collect macrotexture data for state DOTs and private clients. It is noted that several companies have indicated preliminary use of laser systems to collect “3D” macrotexture profiles (the profiles are gathered at a near- continuous level in the transverse direction, but there is a gap between profiles in the longitu- dinal [travel] direction). This is a departure from the legacy method of using a single-spot laser to gather a longitudinal profile that is only one data point wide. However, by using line lasers, the full width of the wheelpath or even the pavement lane can be measured, resulting in a large volume of additional transverse data. This is a method that may merit further investigation. The non-HSLE devices listed in Table 1 are considered devices for reference measurements of pavement macrotexture. Again, the list includes only technologies that either are currently used to gather macrotexture data or at near readiness for commercial application. The devices listed provide profiles of the macrotexture of the pavement’s surface. However, there are still some questions as to whether these measurements are accurate depictions of the actual macrotexture available for tire interaction. To compare readings made at highway speed on the pavement network, reference measurements should be made to establish a basis for comparison. These reference measurements are made under more controlled conditions. Of the equipment listed in Table 1, the most common methods used for reference measurements are the C.T. Meter and the sand patch test. This is for good reason. These two methods have been used extensively in the United States, creating a wealth of data. The results from the C.T. Meter can be expressed as an average depth (MPD) or as a variation from the mean value (RMS), and they have been found to correlate with the results of the sand patch test. 2.2 Available and Emerging Parameters for Characterizing Macrotexture Several parameters have been developed and implemented over the years to convey macro- texture data in a single useful term for further analysis of pavement surface-related phenomena. Company Home Office The Transtec Group Austin, TX Surface Systems and Instruments, LLC (SSI) Mill Valley, CA Pathway Services Tulsa, OK Mandli Communications, Inc. Madison, WI International Cybernetics Corp. Largo, FL Fugro Roadware Mississauga, Ontario, Canada Dynatest Consulting, Inc. Alpharetta, GA Ames Engineering Ames, IA Pavetesting Oxfordshire, UK ARRB Group South Victoria, Australia Ramboll RST Malmö, Sweden VTI Linkoping, Sweden Roadscanners Rovaniemi, Finland Siteco Bologna, Italy VARS BRNO Brno, Czechia Geenwood Engineering Brøndby, Denmark Sineco SpA Milan, Italy Viatech Kongsberg, Norway VECTRA and Lehmann Partner Paris, France Table 5. Macrotexture measurement service providers (as of 2018).

Macrotexture Measurement Technologies and Parameters 29   2.2.1 Commonly Used Macrotexture Characterization Parameters Table 6 summarizes the most common macrotexture parameters and their constituents in use today. This section provides explanations of each parameter or group. MTD Any reference to “texture depth” refers to the volume of void space below the peak(s) of the pavement texture down to the overall pavement surface sampled. The parameter mean texture depth (MTD) is computed from the measurements made for the sand patch test as described in ASTM E965 (2015). In this test, a known volume of sand is spread over the surface of the pavement to create a circular surface. The volume of the sand is divided by the area required to spread the sand resulting in a 1-D measure of texture (thickness) to describe the 3D void volume tested. This test is very time and labor intensive, requires traffic control, and is affected by many operational conditions (e.g., wind, the person performing the test). As this study is related to network-level macrotexture measurements, the sand patch test (or any other physical media volume-based test) is practical only to crosscheck network-level data or for establishing ground truth readings. MPD Mean profile depth (MPD) has become the de facto standard for characterizing network-level macrotexture in the United States. It is essentially the average height of the two highest peaks in an equally divided 100 mm long segment above the mean height of the segment. This parameter is only applicable for 2D profiles; however, 3D profiles can recreate it for any given continuous 2D data sampled from the 3D data. Because it is a measure of the two highest peaks in a profile, MPD is sensitive to outlier peaks. ETD An estimated texture depth (ETD) is used when the MTD is estimated from the MPD using the transformative equation provided in ASTM E1845 (2015). The standard goes on to warn the user that “ETD values should be similar to MTD values calculated via a volumetric technique; however, differences can exist.” This relationship is useful to compare MPD values computed Parameter Reference Strengths Limitations Mean Texture Depth (MTD) ASTM E965 (2015) Time-tested Operator error Mean Profile Depth (MPD) ASTM E1845 (2015), ISO 13473-1 (2019) Time-tested and widely used Only measures a line along the roadway Estimated Texture Depth (ETD) ASTM E1845 (2015), ISO 13473-1 (2019) Relation to MTD Collected by MPD equipment A correlation-based parameter Sensor -Measured Texture Depth (SMTD) Roe et al. (1998) Uses a statistical measure (vs. the MPD’s two peaks) Only measures a line along the roadway Profile Depth (PD) ASTM E1845 (2015) A basic measure; information can be further processed Uses a single peak height as a reference Texture Depth (TD) ISO 13473-1 (2019) A basic measure; information can be further processed Uses the average of the three highest peaks in a 3D profile Root Mean Square (RMS) Wennink and Gerritsen (2000) A stronger statistical basis; describes variation Not widely used in the United States Texture Spectra Power Spectral Density (PSD) Texture Power Spectra Texture Level (TL) Goubert (2007), Anfosso-Lédée and Do (2002), Leandri and Losa (2015) Relation to road noise; some operations are computationally simpler Not widely used in the United States; can require additional analysis Table 6. Available parameters for characterizing macrotexture.

30 Protocols for Network-Level Macrotexture Measurement from macrotexture data with known MTD values (i.e., from sand patch tests) for the pavement and vice versa. However, it has been shown that the equation is not valid for all types of surfaces. The transformative equation given in ASTM E1845 (2015) is: [ ]( )= +0.2 0.8 , (5)ETD MPD mm where MPD and ETD are expressed in mm. RMS Outside of the United States, the root mean square (RMS) is a very common parameter used to characterize macrotexture. RMS is the square root of the mean of the squares of a macro- texture profile. It is a measure of the variation of measurements in a dataset. Since the calculation involves the square root of the summation of the variation, it is a standard deviation of the data. “Large variations” mean large deviations from the mean texture level. The RMS can be used to determine the macrotexture orientation of a pavement by taking the ratio of MPD to RMS according to the following equation: ( ) = . (6)Texture Ratio TR MPD RMS Typically, texture ratios greater than 1.05 indicate a positive macrotexture (more asperities above the mean macrotexture level) while a ratio of less than 0.95 indicates a negative macro- texture (greater void spaces below the mean macrotexture line). SMTD The sensor-measured texture depth (SMTD) is a common parameter computed from the macrotexture data collected by equipment in the European Union and Great Britain. SMTD uses a statistical measure, the RMS, to describe the macrotexture above or below the mean macrotexture level. Viner et al. (2006) stated that the correlation between the MPD and the SMTD depends on the type of surface evaluated, making the arguments for or against either method “finely balanced.” Viner et al. (2006) provide the following relationship between the two parameters: = ×1.42 , (7)0.840MPD SMTD where MPD and SMTD are expressed in mm. PD ASTM E1845 (2015) defines profile depth (PD) as the “difference between the amplitude measurements of pavement macrotexture and a horizontal line through the top of the high- est peak within a given base length.” This definition is similar to the definition that is given in ISO 13473-1 (1997), except the latter indicates the PD is measured over a distance “in the same order of length as that of a car tire/pavement interface” as opposed to the ASTM’s constant 100 mm base length. This measure, on its own, is not necessarily a parameter typically reported for network-level macrotexture data. However, it is fundamental to the calculation of the MPD, and it can be used as a basic measure of pavement surface amplitude at any point relative to the highest peak evaluated.

Macrotexture Measurement Technologies and Parameters 31   TD Texture depth (TD) is typically a 3D parameter. ISO 13473-1 (1997) defines TD by stating “in the three-dimensional case, the distance between the surface and a plane through the top of the three highest peaks within a surface area in the same order of a size as that of a car tire/ pavement interface.” Texture Spectra Texture spectra are a family of parameters that characterize a pavement’s macrotexture by analysis of all the different (spectra) wavelengths that comprise the surface. In the case of pave- ments, a wavelength can be the exposed profile of a single large aggregate, the “trough” formed by the void space between two aggregates, etc. This analysis is typically accomplished by con- verting a signal from the spatial (time) domain to the frequency domain by means of the Fourier transform (or, because the situation involves a finite data set, a discrete Fourier transform). This classification of parameters is useful for identifying repeating patterns in texture, characterizing spatial components (i.e., length, spacing, etc.), and comparison to other surface phenomena such as tire-pavement noise. Examples involving spectral analysis include: • Power Spectral Density (PSD). This is the most common form of spectral analysis performed on macrotexture data collected on a pavement. The premise is the representation of how the power (some measure of the amplitudes) of the signal (i.e., 2D profile gathered with contact- less sensor) is distributed over the frequencies evaluated. This is accomplished by normalizing the power of the signal by the bandwidth of the evaluated frequencies. • Texture Power Spectra. Use of the Fourier transform to convert raw macrotexture profile data into spatial wavelengths, often grouped in octave (i.e., 1, 2, 4, 8) or third-octave bands (Anfosso-Lédée and Do 2002) to perform acoustical analysis due to the similarities of these bands to auditory responses. • Texture Level. Representation of Fourier-transformed profiles in decibels (dB), often used to compare texture measurements to tire-pavement noise generation (Goubert 2007). The texture level (Lx) is computed via the following equation: 20 , (8)10=      L log a a x x ref where ax = profile RMS (m), aref = (−106 m) = reference RMS value of the texture profile amplitude, and x = subscript indicating a value obtained with a certain filter. 2.2.2 Emerging Macrotexture Parameters Table 7 summarizes some emerging macrotexture parameters, and the following sections provide brief explanations of each parameter or group. For each of the 3D parameters, current line-laser technology for high-speed macrotexture measurement devices does not provide a high enough sampling rate for 3D measurements and so these parameters were not considered in the research. MTD3 Liu et al. (2016) used a stationary 3D line-laser scanner to collect high-resolution (< 0.05 mm vertical resolution with < 0.05 mm sampling interval) 3D scans of pavement texture. A new parameter, the digitally simulated 3D MTD (MTD3), was suggested as an analog to the MTD from the sand patch test. Several different technologies (e.g., Pavemetrics LCMS) can be used to collect 3D data on which the “digital sand patch method” can be simulated.

32 Protocols for Network-Level Macrotexture Measurement One advantage of using the MTD3 is that the microtexture spectrum of pavement wave- lengths (< 0.5 mm) can be captured given such a device’s resolution. In addition, relations can be made to other macrotexture parameters with extensive existing datasets, such as MTD and MPD. For example, a linear correlation between the MPD and the MTD3 was established with an r2 = 0.94 according to the following equation: = α + β 3. (9)1 1MPD MTD RMSD3 The 3D RMS Deviation (RMSD3) parameter is calculated in much the same way as the RMS for a 2D profile, but it uses data in both transverse and longitudinal directions (or any orienta- tion in between) for the sample area. The relationship between the RMS computed from C.T. Meter data and the RMSD3 calculated through regression parameters α2 and β2 had an r2 as high as 0.98, resulting in the following equation (Liu et al. 2016): 3. (10)2 2 p= α +βRMS RMSD MPDi The mean profile depth from 3D image measurements (MPDi) parameter arose due to an experiment carried out by El Gendy and Shalaby (2007) in which a four-source photometric Parameter Reference Strength Limitation Digitally Simulated 3D MTD (MTD3) Liu et al. (2016) High resolution; correlation to established parameters Most often gathered by a stationary device 3D RMS Deviation (RMSD3) Liu et al. (2016) High resolution; correlation to established parameters Most often gathered by a stationary device Mean Profile Depth from 3D image measurements (MPDi) El Gendy and Shalaby (2007) High resolution; correlation to established parameters Typically gathered by a stationary device Enveloping Profiles (Goubert 2007) • Empirical • Physical • Spectral • Effective Area of Water Evacuation (EAWE) Von Meier et al. (1992) Klein et al. (2004) Clapp (1983) Mogrovejo et al. (2016) Accounts for a more realistic area for water evacuation. Improved correlations to friction and noise Not implemented in existing software or measurement schemes Wavelet Transformations Leandri and Losa (2015)Zelelew et al. (2013) Greater granularity on measured profile waveform Processing intensive Huang–Hilbert Transform (HHT) Rado and Kane (2014) Good correlation to friction Limited testing; intense post- processing Summit Analysis Le Gal et al. (2008) In-depth analysis of macrotexture asperities Intense post- processing; limited testing 3D Void Volume Sanders et al. (2014) High-resolution 3Ddata Very sensitive to outliers Geometric Statistical Methods • Avg. Roughness (Ra) • Mean Square Roughness (Rq) • Skewness (Rsk) • Kurtosis (Rku) ISO 4288 (1996) ISO 4287 (1997) ASME B46.1 (2009) A common set of tools available to multiple disciplines Not widely used in common pavement surface vernacular Tortuosity Praticò et al. (2017) Use in pervious and porous pavements Difficult to characterize on a network level Rugosity Du Preez (2015) Relates micro and macrotexture Difficult to characterize on a network level Table 7. Summary of emerging macrotexture parameters.

Macrotexture Measurement Technologies and Parameters 33   stereo system was used to produce pavement profiles from 3D images that could be processed into various macrotexture parameters. The work included laboratory sampling of a pavement specimen using both a physical stylus with an in-line dial gauge and a profile gathered from the photography. A relationship between the MPD computed using the data from the digital images and the data from the dial gauge was established. Equation 11 was proposed to relate the MPDd (computed from dial gauge readings) to the MPDi (computed using data from the images): = +0.41 0.21, (11)MPD MPDi d where MPDi and MPDd are expressed in mm. Enveloping Profiles Enveloping profiles seek to accurately model the true tire shape as it passes over the pavement (Goubert 2007). All previous work to characterize a pavement’s macrotexture has been con- cerned with the profile (2D or 3D) of the pavement surface alone or involving a flat plane above the pavement. The underlying assumption is that the void space below the upper plane defined by the parameter will be available for water evacuation below the tire. However, a tire is not a flat infinitely stiff plane. Some of the tire material will deform into the macrotexture’s void space, effectively decreasing the amount of water that can be displaced beneath the tire. The three primary methods used to define the enveloping profile of the tire are as follows: 1. Empirical: Procedures like the Von Meier-Van Blockland-Descornet procedure (Von Meier et al. 1992); 2. Physical: The Hamet-Klein approach (Klein et al. 2004); and 3. Spectral correlation: Spectral correlation of the surface profile (Clapp 1983). An additional parameter used in defining enveloping profiles that shows great promise to more accurately characterize effective pavement macrotexture is the effective area for water evacuation (EAWE) (Mogrovejo et al. 2016). EAWE uses the von Meier et al. method to obtain the enveloped tire profile using the 2D macrotexture data collected from high-speed, single- spot laser equipment. This information is then used to calculate a more realistic area below the enveloped tire profile for water displacement. Comparison of computed EAWE values with MPD values revealed that the MPD tends to overestimate a pavement’s ability to evacuate water. Furthermore, correlations to pavement friction and tire-pavement noise improved when using the EAWE instead of the MPD. Wavelet Transformations A potential group of parameters for analyzing macrotextures can be obtained using wavelet transformations. These parameters rely on a form of signal analysis whereby the signal is broken down into a set of complementary “wavelets” that begin and end at an amplitude of zero and represent a brief oscillation. A primary benefit of wavelet analysis, as found by Zelelew et al. (2013), is that spatial information is maintained when using a C.T. Meter (or other equipment that captures data in multiple directions), whereas spatial information is lost if a simple Fourier transformation is used. Unfortunately, analyzing wavelet transformations is computationally intensive, and this procedure has not been implemented in commercially available equipment for pavement macrotexture analysis. HHT The Hwang–Hilbert Transform (HHT) attempts to simplify other macrotexture frequency analysis methods (i.e., Fourier or wavelet transformations) while simultaneously improving correlations to pavement friction measurements (Rado and Kane 2014). The transformation

34 Protocols for Network-Level Macrotexture Measurement decomposes a macrotexture profile to a limited number of profiles and then produces averaged frequency and amplitude profiles for correlation to pavement friction. The processing is labor intensive and, at least at this time, does not appear practical for implementation on network-level analysis. Summit Analysis This family of analyses closely studies the summits (“peaks of asperities”) of a measured macro texture profile. This information is of great use in studying tire-pavement interaction and has been used by tire developers to improve their products. A full study includes analyzing the summit density, distribution of summit height radius, and slope. It has the advantage that it can shed light on tire contact, tire deformation (which contributes to the hysteresis compo- nent of friction), and vehicle splash and spray. However, it is also very analysis intensive and requires human interpretation on small areas of pavement measured and is, therefore, not conducive to network-level analysis. 3D Void Volume This method is essentially an MTD computed from 3D surface measurements (i.e., laser texture scanner). In this approach, the volume of voids below the highest measured peak is calculated and then divided by the surface area (Sanders et al. 2014). The units of measure are the volume per unit area. Geometric Statistical Methods This family of tools can be used to describe a surface’s macrotexture. These analyses can be applied to both 2D and 3D macrotexture profiles (Ech et al. 2007); however, these methods are not typically employed to describe pavement macrotexture in the United States. The measure- ments themselves describe statistical moments: with each increasing moment, each individual data point is raised to a higher power. For each moment, the distance (y) refers to the distance between the data point and the slope-adjusted, zero-centered mean for the profile evaluated, and the number (n) refers to the number of data points evaluated. Two of the parameters, skewness and kurtosis, are sensitive to outlier peaks (either true outlier peaks or errors) in the data, a limi- tation that is similar to that of the MPD. Four parameters—average roughness (Ra), mean square roughness (Rq), skewness (Rsk), and kurtosis (Rku)—merit discussion here: • Outside the pavement measurement community, average roughness is also referred to as the “mean profile depth.” This first statistical moment is simply the mean of the dataset. In pavement macrotexture terms, the calculation uses the absolute value of the data because the data have been adjusted to a mean of zero, as shown in the following equation: ∑ = =R y n a ii n , (12)1 where y = elevation from slope-suppressed, zero-mean profile segment, and n = number of samples in segment evaluated. The Ra for the example data is shown in Figure 13. • The second statistical moment, mean square roughness, is a description of variance (the width of the distribution of measurements of the dataset). This term is synonymous with the RMS as described in the previous section. ISO 13473-2 (2002) defines mean square rough- ness as follows:

Macrotexture Measurement Technologies and Parameters 35   ∫ ( )= 1 , (13)2 0 R l Z x dxms l where l = length of segment evaluated, and Z = elevation from slope-suppressed, zero-mean profile segment. The mean square roughness for the example data is shown in Figure 14. The equation to calculate this parameter for discrete data can be written as follows: ∑= = . (14) 2 1R y n q ii n Taking into account that a texture ratio can be determined by dividing the MPD by the RMS, the example profile would have a texture ratio of 1.88 ÷ 1.02 = 1.84. Thus, it would generally be regarded as having a positive macrotexture (more upward asperities than voids below the aver- age plane). • Skewness (Rsk), shown in Figure 15, can be used to capture positive or negative macrotexture behavior. Negative skewness values indicate a negative macrotexture (more troughs than peaks). Skewness is calculated as follows: ∑ ( ) = =R y n R sk ii n q . (15) 3 1 3 Figure 13. Average roughness for the example data (0.8277).

36 Protocols for Network-Level Macrotexture Measurement Figure 14. Mean square roughness (1.02) for the example data. Figure 15. Skewness (–0.76) for the example data.

Macrotexture Measurement Technologies and Parameters 37   • Kurtosis (Rku) can generally be used to describe the peakedness of a macrotexture profile (i.e., how severe peaks and troughs are; hence the large value shown in Figure 16, given that the data has not been filtered). Statistical moments of four seldom are used to describe surface characteristics; however, in image processing, kurtosis is sometimes used to describe the uniformity of the grayscale distribution. Kurtosis is calculated as follows: ∑ ( ) = =R y n R ku ii n q . (16) 4 1 4 Tortuosity Tortuosity is a measure of “twistiness.” Typically used to describe porous materials, such as porous friction courses or pervious pavements. The measure is more common in acoustical studies of pavement (Praticò et al. 2017) than for other general pavement surfaces. Typically taken as the ratio of the average length of all the capillaries to the length of the overall porous material (i.e., the thickness of the pavement layer evaluated), the measure could prove useful for measuring the macrotexture of pavements with many interconnected voids. Rugosity Rugosity deals with very fine changes in the amplitude of a profile (Du Preez 2015). Typically taken as the ratio of the true surface area to the geometric surface area (i.e., one converted to geometric shapes as is done in profiling), rugosity can be conceptualized in terms of pavement texture as the ratio between microtexture and macrotexture. Figure 16. Kurtosis (2.38) for example data.

38 Protocols for Network-Level Macrotexture Measurement 2.3 Data Filtering and Correction A basic tenet of signal processing is that all data will have some noise masking the physical phenomenon measured. Non-contact macrotexture measurements are no different. Noise can come from various sources, such as the sensor used (calibration and system checks should be performed before fieldwork to minimize these effects) or from attributes of the surface being measured (i.e., reflective versus non-reflective surfaces, transparency of the surface, etc.). Therefore, one first step before further signal processing occurs is often to filter the data appropriately. For example, ASTM E1845 (2015) requires that “spatial frequency components above 400 cycles per meter that correspond to a texture wavelength of 2.5 mm shall be removed.” Because the “noise” in the signal of the pavement macrotexture for this research was in the shorter wave- lengths, a low-pass filter was typically used on the data before further processing. This meant that attributes of the signal, such as large aggregates whose wavelength is of a lower frequency, passed through the filter, but higher-frequency data (smaller perturbations in the data stream) were filtered out. Figure 17 shows a raw macrotexture profile collected on the Virginia Smart Road that was subsequently filtered using a finite impulse response filter with a cutoff frequency of 40 Hz and a sampling rate of 200 Hz. Of note is the change in the mean value when the data is filtered. Specifically, the mean value for the “slope = zero” line is 0.18 mm less than the mean value of the “mean = 0” line. Because pavements are sloped in the longitudinal and lateral directions and because slopes can be created from small movements of the vehicle suspension, the slope of the surface could be erroneously interpreted as a shallowing or deepening of the macrotexture in analyses of Raw data mean = 19.7, filtered data mean = 19.1. M ea su re m en t M ea su re m en t Location (mm) Location (mm) Figure 17. Data filtering.

Macrotexture Measurement Technologies and Parameters 39   small segments of data. To correct this issue, a linear regression of the data is typically made, with the resulting equation of the line subtracted from the data to suppress the slope and bring the mean value of the profile to zero. By bringing the mean value to zero, any description of the data (e.g., minimums, maximums, averaging of points) will be relative to zero. This yields familiar values, such as MPD, on the order of 1 mm instead of large values relative to some other datum. Figure 18 shows the data from Figure 17 as normalized (using linear regression) to a mean value of zero and an overall slope of zero. Two other vexing problems in macrotexture measurement stem from the phenomena of dropouts and spikes. Such phenomena are represented by sudden outlier peaks (upward for spikes and downward for dropouts) that do not seem to match the rest of the data. Such outliers can be caused by the laser emitted to the surface not returning with sufficient strength (dropout), sudden vertical movement of the vehicle carrying the sensor (movements not adequately elec- tronically dampened by inertial measurement units associated with the device), or reflectivity properties of the material measured (spikes). No matter the cause, the outlier values do not accurately represent the surface of the pavement measured. Various equipment manufacturers have incorporated correction algorithms in their hardware or software to address the issue of dropouts. Data points beyond some threshold are typically flagged by the system at the time of acquisition so they can be subsequently dealt with. Figure 19 and Figure 20 show the impact of a single (fictitious) outlier on the data. Note the change in the mean value, which would errone- ously result in a higher or lower MPD for the segment. Schemes to deal with the presence of spikes and dropouts typically look at the data stream for received values that are out of the bounds set by the sensor manufacturer. Dropouts are flagged and typically thrown out, replaced by interpolated values of valid points before and after the Figure 18. Use of linear regression to bring data to mean of zero (blue line) and suppress slope (orange line).

40 Protocols for Network-Level Macrotexture Measurement Filtered data mean = 19.1, mean with dropout = 19.0. M ea su re m en t ( m m ) M ea su re m en t ( m m ) Location (mm) Location (mm) Figure 19. Effect of dropout on data. Filtered data mean = 19.1, mean with spike = 19.0. M ea su re m en t ( m m ) M ea su re m en t ( m m ) Location (mm) Location (mm) Figure 20. Effect of spike on data.

Macrotexture Measurement Technologies and Parameters 41   outlier. For example, ISO 13473-1 (1997) recommends the following approach to interpolate the replaced value: ( )= − − − + , (17)z z z n m i m zi n m m where i = the sample number where the value is invalid, m = the sample number of the nearest valid value before i, n = sample number of the nearest valid value after i, zi = the interpolated value for sample i, zm = the value of sample m, and zn = the value of sample n. Identification of outliers can also be achieved via statistical means. However, Katicha et al. (2015) found that procedures such as looking for values outside of a certain number of standard deviations or the Bonferroni method could often miss outlier data. Instead, the application of the false discovery rate method and a family of generalized Gaussian distributions performed much better at removing nearly all spikes. Because this method adapts itself to the incoming data instead of user-selected values that may be inappropriate as the characteristics of the pave- ment change, it is very attractive for evaluation of network-level macrotexture measurements.

Next: Chapter 3 - Data Collection »
Protocols for Network-Level Macrotexture Measurement Get This Book
×
MyNAP members save 10% online.
Login or Register to save!
Download Free PDF

Macrotexture, which influences vehicle-roadway skid resistance, refers to the texture of the pavement due to the arrangement of aggregate particles. Pavement surfaces are subjected to seasonal variations, and over time the embedded aggregates become polished due to the onslaught of traffic. Research has shown that wet-weather crashes are influenced by the macrotexture of the pavement surface.

The TRB National Cooperative Highway Research Program's NCHRP Research Report 964: Protocols for Network-Level Macrotexture Measurement provides state transportation pavement engineers and other practitioners with recommended protocols for macrotexture test measures, equipment specifications, and data quality assurance practices.

  1. ×

    Welcome to OpenBook!

    You're looking at OpenBook, NAP.edu's online reading room since 1999. Based on feedback from you, our users, we've made some improvements that make it easier than ever to read thousands of publications on our website.

    Do you want to take a quick tour of the OpenBook's features?

    No Thanks Take a Tour »
  2. ×

    Show this book's table of contents, where you can jump to any chapter by name.

    « Back Next »
  3. ×

    ...or use these buttons to go back to the previous chapter or skip to the next one.

    « Back Next »
  4. ×

    Jump up to the previous page or down to the next one. Also, you can type in a page number and press Enter to go directly to that page in the book.

    « Back Next »
  5. ×

    To search the entire text of this book, type in your search term here and press Enter.

    « Back Next »
  6. ×

    Share a link to this book page on your preferred social network or via email.

    « Back Next »
  7. ×

    View our suggested citation for this chapter.

    « Back Next »
  8. ×

    Ready to take your reading offline? Click here to buy this book in print or download it as a free PDF, if available.

    « Back Next »
Stay Connected!