repeatability and reproducibility of mobile retroreflectivity units for

16
REPEATABILITY AND REPRODUCIBILITY OF MOBILE RETROREFLECTIVITY UNITS 1 FOR MEASUREMENT OF PAVEMENT MARKINGS 2 3 4 Bouzid Choubane 1 (Corresponding Author), Joshua Sevearance 2 , Hyung Suk Lee 1 , Patrick 5 Upshaw 1 , and James Fletcher 2 6 7 8 9 (1) Florida Dept. of Transportation, Materials Research Park 10 5007 N.E. 39th Avenue, Gainesville, FL 32609 11 Phone: (352) 955-6341 12 Fax: (352) 955-6345 13 Email: [email protected] 14 Email: [email protected] 15 Email: [email protected] 16 17 (2) University of North Florida (UNF), Mechanical Engineering 18 1 UNF Drive, Jacksonville, FL 32224 19 Phone: (904) 620-1844 20 Fax: (904) 620-1391 21 Email: [email protected] 22 Email: [email protected] 23 24 25 26 27 28 29 30 31 32 33 34 SUBMITTED FOR PRESENTATION AT THE 2013 TRANSPORTATION RESEARCH BOARD 35 MEETING AND PUBLICATION IN THE TRANSPORTATION RESEARCH RECORD 36 37 38 39 40 41 Word Count 42 43 Abstract = 248 44 Body Text = 3,352 45 Tables 1 x 250 = 250 46 Figures 11 x 250 = 2,750 47 Total = 6,600 48 49 TRB 2013 Annual Meeting Paper revised from original submittal.

Upload: truongdan

Post on 13-Feb-2017

221 views

Category:

Documents


2 download

TRANSCRIPT

REPEATABILITY AND REPRODUCIBILITY OF MOBILE RETROREFLECTIVITY UNITS 1 FOR MEASUREMENT OF PAVEMENT MARKINGS 2

3 4

Bouzid Choubane1 (Corresponding Author), Joshua Sevearance2, Hyung Suk Lee1, Patrick 5 Upshaw1, and James Fletcher2 6

7 8 9

(1) Florida Dept. of Transportation, Materials Research Park 10 5007 N.E. 39th Avenue, Gainesville, FL 32609 11

Phone: (352) 955-6341 12 Fax: (352) 955-6345 13

Email: [email protected] 14 Email: [email protected] 15

Email: [email protected] 16 17

(2) University of North Florida (UNF), Mechanical Engineering 18 1 UNF Drive, Jacksonville, FL 32224 19

Phone: (904) 620-1844 20 Fax: (904) 620-1391 21

Email: [email protected] 22 Email: [email protected] 23

24 25

26 27 28 29 30 31 32 33 34 SUBMITTED FOR PRESENTATION AT THE 2013 TRANSPORTATION RESEARCH BOARD 35

MEETING AND PUBLICATION IN THE TRANSPORTATION RESEARCH RECORD 36 37 38 39 40 41 Word Count 42 43 Abstract = 248 44 Body Text = 3,352 45 Tables 1 x 250 = 250 46 Figures 11 x 250 = 2,750 47 Total = 6,600 48 49

TRB 2013 Annual Meeting Paper revised from original submittal.

Choubane, Sevearance, Lee, Upshaw, and Fletcher 1

ABSTRACT 1 2 The Florida Department of Transportation (FDOT) has historically used a combination of handheld 3 devices and visual surveys to evaluate the retroreflectivity of pavement markings. However, visual 4 surveys have the inherent limitations of operator bias while the use of a handheld device is slow, labor 5 intensive, and presents safety hazards. Many highway agencies have recognized that a Mobile 6 Retroreflectivity Unit (MRU) may be a safer and a more efficient alternative to the handheld 7 retroreflectometers. Since the measurement process relies on the operator-driven instrument, a level of 8 uncertainty is always a concern in evaluating pavement markings with the MRU. This research is aimed 9 at assessing the precision and bias of the MRU while using the handheld retroreflectometer as a reference 10 device. A total of ten 1.0 mile long field sites were selected to include various pavement surface types 11 and pavement marking materials (paints and thermoplastics). The results indicated that, when compared 12 to the handheld retroreflectometers, the MRU demonstrated no statistical differences or bias at a 95 13 percent confidence level for the retroreflectivity values ranging between 200 and 800 mcd/m2/lux. In 14 addition, it was determined that the retroreflectivity values from two properly conducted tests using a 15 single MRU on the same pavement marking should not differ by more than 7.8 percent and when 16 different MRUs are used on the same pavement marking, the retroreflectivity values should not differ 17 more than 13.3 percent. This paper presents a description of the testing program, the data collection effort, 18 and the subsequent analyses and findings. 19

20

TRB 2013 Annual Meeting Paper revised from original submittal.

Choubane, Sevearance, Lee, Upshaw, and Fletcher 2

INTRODUCTION 1 2 Retroreflectivity of pavement marking is one of the key factors for night time visibility, safety, and 3 comfort to those traveling on the state highway network. For this reason, the level of retroreflectance 4 provided by the pavement marking needs to be monitored accurately and maintained appropriately. In 5 recognition of the importance of the pavement marking retroreflectivity, the Federal Highway 6 Administration (FHWA) released their Docket number FHWA-2009-0139 in April, 2010 which proposed 7 that the Manual on Uniform Traffic Control Devices (MUTCD) include standards, guidance, options, and 8 supporting information needed for maintaining the minimum levels of pavement marking retroreflectivity 9 (1). As a result, the proposed revisions to the MUTCD including the minimum levels of pavement 10 marking retroreflectivity has been designated as Revision 1 to the 2009 Edition of the MUTCD (2). 11 12

Historically, the Florida Department of Transportation (FDOT) has used a combination of 13 handheld retroreflectometers and visual surveys to evaluate the retroreflectivity of pavement markings. 14 However, the disadvantage of the handheld units is that the data can only be collected at discrete 15 locations. In addition, testing with the handheld devices is labor intensive and presents potentially 16 hazardous to the operator and to the traveling public as testing requires maintenance of traffic. 17 Furthermore, visual surveys have the inherent limitations of operator bias due to subjective opinions 18 which make quantifying retroreflectance near impossible. 19

20 The Mobile Retroreflectivity Unit (MRU) was developed to measure retroreflectivity of 21

pavement markings continuously at highway speeds. Due to its improved safety and efficiency, MRU has 22 been recognized by many highway agencies as an appealing alternative to the handheld 23 retroreflectometers. The primary benefit of the MRU is that it offers a more efficient and objective means 24 for obtaining the inventory data needed for maintaining the minimum level of pavement marking 25 retroreflectivity provided in the MUTCD. 26

27 28 BACKGROUND 29 30 FDOT, with technical support provided by the University of North Florida (UNF), initiated a program to 31 first evaluate and optimize the MRU performance followed by a comprehensive implementation plan for 32 the MRU as an alternative for evaluating pavement marking retroreflectance. 33 34 Evaluation Process 35 36 The first step in the evaluation process was to complete a literature review, including a comprehensive 37 discussion with most of the MRU users in the United States. There was a wide variety of experiences, but 38 most users were limited in the implementation due to shortcomings with the device (3). Many of the 39 MRU users reported calibrating the MRU often (up to 20 times a day) in order to achieve reliable results. 40 Other operators discussed efforts to maintain the “30 meter geometry” which is critical to accurate MRU 41 measurement. Road conditions (hills, curves, etc.), changes in vehicle conditions (driver, fuel level, etc.), 42 and vehicle dynamics (pitch, roll, acceleration/deceleration, etc.) were all identified as culprits resulting in 43 difficulty in obtaining accurate and repeatable MRU measurement (3, 4). 44 45 MRU Optimization 46 47 The initial effort was to develop and implement mitigation strategies to improve the MRU performance. 48 For example, early testing showed that measurements taken within a short period of time were repeatable, 49 but over a course of a day the results varied significantly. It was determined the MRU had a significant 50 temperature sensitivity, specifically the interference filters used to eliminate background lighting. To 51

TRB 2013 Annual Meeting Paper revised from original submittal.

Choubane, Sevearance, Lee, Upshaw, and Fletcher 3

mitigate this issue, a thermoelectric cooler was installed to maintain the operating temperature and 1 temperature compensation was added to the software (5). 2 3

Through testing, it became clear that the critical step in order to obtain accurate MRU 4 measurement was the calibration process. The consistency of the field data and the accuracy of the MRU 5 measurements are highly dependent on the procedure and the material used for its calibration. Calibration 6 standards, consisting of a short section of beaded line striping are similar material to pavement markings 7 but are very difficult to accurately measure (6, 7, 8, 9). Much effort has been spent to improve and 8 optimize the calibration procedure and alternative calibration tools (6, 7). 9 10 Precision and Bias of MRU 11 12 FDOT has been researching the implementation of the MRU for evaluating pavement markings. The 13 benefits of implementing the MRU to evaluate line-marking retroreflectivity include optimization of 14 maintenance efforts, evaluation of construction techniques and new products, and improvement of the 15 overall safety of the driving public. The preliminary goals of the implementation plan include utilizing the 16 MRU to identify pavement markings that are approaching the minimum allowable retroreflectivity, 17 monitoring retroreflective characteristics of various materials/applications of pavement markings, and 18 providing inventory for maintenance to assess restriping strategies. 19 20

However, in order for FDOT to survey the statewide pavement markings, it was necessary to 21 have multiple MRUs and operators available. Due to the operator subjectivity and the computational 22 components involved with retroreflective measurements, assessing the precision of the MRU became 23 critical as part of the implementation effort. Hence, FDOT conducted a study in 2010 to assess the 24 precision of the MRU in terms of repeatability, but this preliminary effort was limited to a single MRU 25 (10). The follow up study, which is the basis of this paper, was initiated to assess the precision of the 26 MRU in terms of repeatability and reproducibility for pavement marking retroreflectivity. Handheld 27 retroreflectometers were used as a reference device in assessing the MRU’s precision. This paper 28 documents the test protocols, results, and findings of the study. 29 30 31 OBJECTIVES 32 33 The primary objective of this study was to assess the precision and bias of the MRU for determining the 34 retroreflective characteristics of in-service pavement markings in Florida. The precision of the MRU was 35 expressed in terms of repeatability and reproducibility while the bias was evaluated using the handheld 36 retroreflectometer as a reference device. 37 38 39 RETROREFLECTANCE 40 41 Retroreflectivity of pavement markings and traffic signs is an important part of roadway guidance and 42 safety, especially at night. Pavement markings reflect light from the vehicle’s headlamps back to the 43 operator’s eyes. This process is called Retroreflectance (RL), and is quantified as the ratio of the 44 luminance (or brightness) of an object as detected by a sensor to the illuminance of the object by a light 45 source and is expressed in units of millicandelas per meter squared per lux (mcd/m2/lux). Pavement 46 markings typically provide retroreflectivity through the application of small glass spheres (commonly 47 called beads) that are partially embedded into the pavement marking material. This allows incoming light 48 from the vehicle headlamps to reflect back to the origin of the light source, as illustrated in Figure 1. 49 50

TRB 2013 Annual Meeting Paper revised from original submittal.

Choubane, Sevearance, Lee, Upshaw, and Fletcher 4

1 FIGURE 1 Method of creating a retroreflective effect using glass beads 2

3 Pavement markings can consist of various materials such as paints, thermoplastics and tapes, 4

many using the application of glass beads that can vary in size as well. Obtaining a single RL value for a 5 pavement marking using the handheld retroreflectometer or the MRU can be difficult due to factors that 6 influence the amount of retroreflectivity produced by glass beads such as the dispersion in a non-uniform 7 pattern, embedment depth, refraction index, size, clarity, and roundness. In addition, climate conditions 8 such as rain, fog, snow, ultraviolet light and heat can all affect retroreflective properties (11). 9 10 11 TESTING EQUIPMENT 12 13 The equipment used in this study included two MRUs and three handheld retroreflectometers, owned by 14 FDOT. The handheld and mobile retroreflectometers measure the retroreflectance by applying the “30 15 meter geometry” described in ASTM E 1710 (12). The 30 meter geometry consists of the following 16 assumptions: a typical passenger vehicle headlamp height of 0.65 m (2.1 ft.), a driver eye height of 1.2 m 17 (3.9 ft.), and a distance of 30 m (98 ft.) between the headlamps and the ground-based retroreflectance 18 target. In order to reduce the size of the measuring device, the MRU uses a 1/3rd scale of the 30 meter 19 geometry, as shown in Figure 2. The geometry of the handheld retroreflectometer corresponds to a much 20 smaller scale. 21

22

TRB 2013 Annual Meeting Paper revised from original submittal.

Choubane, Sevearance, Lee, Upshaw, and Fletcher 5

1 FIGURE 2 Standard 30 meter geometry and the 1/3 scale used in MRU 2

3 Handheld Retroreflectometer 4 5 The handheld devices used for this study were one Road Vista StripeMaster and two Delta LTL-X 6 retroreflectometers, all conforming to ASTM E 1710 and are shown in Figure 3. The device outputs a 7 digital readout of the measured retroreflectivity to the nearest 1 mcd/m2/lux. The handheld 8 retroreflectometers have supports that are approximately 10.0 cm (4 inches) wide for better stability 9 positioned on the pavement marking and when placed down, are centered in the pavement marking. 10 11

12 FIGURE 3 FDOT’s Handheld Retroreflectometers 13 (StripeMaster on the left and LTL-X on the right) 14

15 Mobile Retroreflectivity Unit 16 17 The two MRUs used in this study were both Road Vista Laserlux retroreflectometers mounted on full-size 18 passenger vans equipped with all the mechanical and electrical power supplies for evaluating pavement 19 markings. In addition, the vehicles include a data acquisition system for collecting and storing 20 information. A distance-measuring instrument (DMI) is provided to determine the position along the 21 roadway. The longitudinal distance measurement is critical in associating the precise location for each 22

TRB 2013 Annual Meeting Paper revised from original submittal.

Choubane, Sevearance, Lee, Upshaw, and Fletcher 6

0.16 km (0.1 mile), the interval at which the MRU is reporting the data. One of the MRUs used in the 1 study is shown in Figure 4. 2 3

4 FIGURE 4 FDOT’s Mobile Retroreflectivity Unit 5

6 The MRU is equipped with a 10 mW Helium Neon (HeNe) laser to provide a controlled light 7

source with a wavelength of 632.8 nm (0.025 mils) (13). The other principal sub-system is the photo 8 detector which is a device that houses a silicon photodiode to convert the retroreflected light into an 9 electrical signal and an amplifier that magnifies the low-level signal to a measurable level (5). 10 Interference filters are utilized to filter out light of wavelengths other than the laser light as the reflected 11 light may consist of wavelengths resulting from other sources such as sunlight, street lighting, vehicle 12 headlamps, etc. A double-sided mirror rotates at 10 Hz and collects the data at a rate of 200 13 samples/scan, creating a scan width of up to 1.1 m (3.6 ft.) wide as the laser and the photodetector sweep 14 the pavement marking (13). Other mirrors are utilized to re-direct the light to achieve the reduced 30 15 meter geometry. Figure 5 is a schematic of the internal components of the retroreflectometer unit and the 16 path of the laser light (11). 17

18

19 FIGURE 5 Internal components and laser path of retroreflectometer 20

TRB 2013 Annual Meeting Paper revised from original submittal.

Choubane, Sevearance, Lee, Upshaw, and Fletcher 7

1 The MRU is a sensitive instrument and the precision of the MRU measurements is highly 2

dependent on the calibration of the equipment. It is important that accurate geometry and an appropriate 3 calibration panel be used during MRU calibration to reduce uncertainty in pavement marking 4 measurements. Any errors introduced during the calibration process will be transferred to errors in 5 pavement marking evaluation. To ensure precise measurement, calibration of the MRU was performed in 6 close proximity to the pavement marking test section, as shown in Figure 6. 7

8

9 FIGURE 6 Calibration of FDOT’s Mobile Retroreflectivity Unit in the field 10

11 The goal of the calibration process is to standardize the MRU by scanning a calibration panel of 12

which the RL value is known. The consistency of the calibration panel is crucial to the success of the 13 MRU calibration and should result in reproducible calibration factors. Since there is no standard 14 procedure broadly accepted for the calibration and operation of the MRU, FDOT has researched and 15 implemented a new calibration panel that has provided more reproducible calibrations. The new 16 calibration panel used by FDOT is a 152 mm (6 inches) wide black reflective vinyl with a RL of 895 17 mcd/m2/lux. A picture of the new calibration panel is shown in Figure 7. For this study, the 18 aforementioned retroreflective panel was consistently used to calibrate the MRU prior to any data 19 collection. 20 21

22 FIGURE 7 FDOT’s Calibration Panel for the Mobile Retroreflectivity Unit 23

24 25

TRB 2013 Annual Meeting Paper revised from original submittal.

Choubane, Sevearance, Lee, Upshaw, and Fletcher 8

DATA COLLECTION 1 2 Ten 1.6 km (1.0 mile) test sections with various pavement surface types (open and dense graded) and 3 pavement marking materials (paints and thermoplastics) were selected within Alachua County, Florida. 4 The test sites were selected to avoid breaks in the pavement marking and to minimize roadway geometric 5 variables such as inclines, declines, and curves. All tests were performed on the 152 mm (6 in.) wide 6 white edge-line. Prior to testing, the beginning and ending limits of the test sections were clearly 7 identified to ensure an accurate point of reference between all MRU tests. 8 9

Three handheld retroreflectometers were initially used to measure the retroreflectivity every 36.6 10 m (120 ft.), resulting in 44 measurements over the 1.6 km (1.0 mile) distance. At each site, the three 11 handheld devices were calibrated and used to measure the retroreflectivity of the pavement marking. The 12 resulting measurements were then averaged for each 0.16 km (0.10 mile) section for direct comparison 13 with the MRU data output. The longitudinal distance between each test was measured using a digital 14 measuring wheel with a 2.5 mm (0.1 in.) resolution. 15

16 Once the handheld measurements were completed, each MRU performed three repeat runs over 17

each test section. The MRUs were aligned in the center of the edge-line and the vehicle wander was 18 minimized to ± 6 inches from the center of the pavement marking. In addition, the same operators were 19 utilized throughout the series of tests and each operated the same MRU. The number of data scans taken 20 by the MRU depends on the travelling speed of the vehicle. On average, FDOT’s MRUs collect 21 approximately 145 scans per 0.16 km (0.1 mile), when traveled at a speed of 80 km/h (50 mph). The 22 MRUs were calibrated prior to taking measurements at each test site. The MRU data was averaged for 23 every 0.16 km (0.1 mile) segment, producing 10 averaged retroreflectivity values per site. 24

25 For the test program, a total of 440 retroreflective measurements were collected for each of the 26

three handheld retroreflectometers. For each of the two MRUs, three runs were performed at each of the 27 ten sites, resulting in a total of 600 average retroreflectivity MRU measurements. 28

29 30

PRECISION AND BIAS ESTIMATES 31 32 Accuracy and precision are two of the most important criteria for the usefulness of any reliable testing 33 device. ASTM E 177 indicates that the accuracy is typically stated in terms of bias which is defined as 34 the difference of the measured values and the accepted reference value (14). It also states that the 35 precision is typically stated in terms of repeatability (within MRU precision) or reproducibility (between 36 MRUs precision). 37 38

ASTM C 670 specifies that the existence of bias should be determined by comparing the t statistic 39 calculated from the difference between the reference and the estimated values to the critical t statistic, tcrit, 40 for a given confidence level (15). For the evaluation of the MRU, the t statistic was calculated based on 41 the difference between the MRU measurements and the reference handheld retroreflectometers. 42

43 In addition, ASTM C 670 states that an “acceptable difference between two test measurements” 44

or the “difference two-sigma limit” (d2s), can be selected as an appropriate index of precision. The d2s 45 index for a 95 percent confidence level can be calculated by multiplying the appropriate standard 46 deviation or coefficient of variation (COV) by 22 . (15). The appropriate standard deviation and 47 coefficient of variation (COV) are those that represent the within and between unit variation due to the 48 multiple MRU measurements made by two operators and two units. In this study, the above statistics 49 were first obtained for each segment, and then pooled to result in an overall estimate of the within unit 50

TRB 2013 Annual Meeting Paper revised from original submittal.

Choubane, Sevearance, Lee, Upshaw, and Fletcher 9

(repeatability) and between unit (reproducibility) variation as outlined in ASTM C 802 (16). The 1 precision statement was then determined based on the pooled statistics. The MRU data as well as the 2 analysis and the resulting precision statements are presented in the subsequent sections. 3

4 STATISTICAL ANALYSIS 5 6 The range and variation in data collected with the three handheld retroreflectometers and two MRUs for 7 each test section are illustrated in Figure 8. Based on the measurements of both devices, the magnitude 8 ranged from 151 to 834 mcd/m2/lux. 9 10

0

100

200

300

400

500

600

700

800

900

0 1 2 3 4 5

Retroreflectance (mcd/m

^2/lux)

Segments for Sites 1 ‐ 5

MRUs

Handhelds

11 (a) Sites 1 - 5 12

0

100

200

300

400

500

600

700

800

900

5 6 7 8 9 10

Retroreflectance (mcd/m

^2/lux)

Segments for Sites 6 ‐ 10

MRUs

Handhelds

13 (b) Sites 6 - 10 14

FIGURE 8 Range of the Retroreflective Data. 15

TRB 2013 Annual Meeting Paper revised from original submittal.

Choubane, Sevearance, Lee, Upshaw, and Fletcher 10

1 The retroreflectivity values for the handheld retroreflectometers were compared to the MRUs and are 2 illustrated in Figure 9. Statistical analysis was also performed to assess the precision of the MRU in terms 3 of bias defined as the systematic error that contributes to the difference between the mean of the MRU 4 and the accepted reference measurement, which in this case is the average of the handheld measurements. 5 A matched-pairs t-test was conducted to test the significance in the mean difference between manual and 6 automated faulting measurements (15). The t and tcrit statistics were calculated as 1.4 and 2.0, 7 respectively. Because the calculated t statistic falls inside of the ±tcrit range, it can be concluded that the 8 MRU does not exhibit bias when compared to the handheld devices. 9 10

y = 0.96x + 7.01

R² = 0.80

0

100

200

300

400

500

600

700

800

900

0 200 400 600 800

Handheld Retroreflectivity M

easurements 

(mcd/m

^2/lux)

MRU Retroreflectivity Measurements (mcd/m^2/lux)

11 FIGURE 9 Handheld retroreflectometer vs. MRU 12

13 ASTM C 802 also states that the form of the precision statement should be determined based on the 14 relationship between the average and the standard deviation of the measurements (16). In order to 15 determine if the repeatability and reproducibility of the MRU are dependent on the level of 16 retroreflectance, the pooled standard deviation and coefficient of variance (COV) were plotted against the 17 average retroreflectivity values and are shown Figures 10 and 11, respectively. As indicated by the slope 18 of the linear regression lines being much less than 1.0 and close to zero, there is no direct relationship 19 between the variability of the collected data and the magnitude of the retroreflectivity values. This 20 indicates that both the repeatability and reproducibility statements could be drawn independent of the 21 retroreflectivity values considered herein. 22 23

24

TRB 2013 Annual Meeting Paper revised from original submittal.

Choubane, Sevearance, Lee, Upshaw, and Fletcher 11

y = 0.02x + 0.82R² = 0.32

0.0

5.0

10.0

15.0

20.0

25.0

30.0

35.0

0 200 400 600 800 1000

Stan

dard Deviation of R

etroreflectivity 

Measurements (mcd/m

^2/lux)

Average Retroreflectivity Measurements (mcd/m^2/lux)

1 a) Within MRU 2

3

y = 0.01x + 12.75R² = 0.03

0.0

5.0

10.0

15.0

20.0

25.0

30.0

35.0

40.0

45.0

50.0

0 200 400 600 800 1000

Standard Deviation of Retroreflectivity 

Measurements (m

cd/m

^2/lux)

Average Retroreflectivity Measurements (mcd/m^2/lux)

4 b) Between MRUs 5

6 FIGURE 10 Standard deviation vs. average retroreflectivity 7

8

TRB 2013 Annual Meeting Paper revised from original submittal.

Choubane, Sevearance, Lee, Upshaw, and Fletcher 12

y = ‐0.00x + 3.36R² = 0.02

0.0

1.0

2.0

3.0

4.0

5.0

6.0

7.0

8.0

0 200 400 600 800 1000

Coefficien

t of Variation of Retroreflectivity 

Measurements (%)

Average Retroreflectivity Measurements (mcd/m^2/lux)

1 a) Within MRU 2

3

y = ‐0.01x + 7.60

R² = 0.15

0.0

2.0

4.0

6.0

8.0

10.0

12.0

14.0

0 200 400 600 800 1000

Coefficient of Variation of Retroreflectivity 

Measurements (%)

Average Retroreflectivity Measurements (mcd/m^2/lux)

4 b) Between MRUs 5

6 FIGURE 11 Coefficient of variation vs. average retroreflectivity 7

8 In addition, the pooled variance, standard deviations, COV, d2s, and d2s% were calculated from the data 9 for the repeatability and reproducibility assessments of the MRU. A summary of the results is shown in 10 Table 1. The table shows that the overall pooled standard deviations of the retroreflectivity for the MRU 11 were calculated to be 12.0 mcd/m2/lux (within MRU) and 18.8 mcd/m2/lux (between MRUs), 12 respectively. 13

TRB 2013 Annual Meeting Paper revised from original submittal.

Choubane, Sevearance, Lee, Upshaw, and Fletcher 13

TABLE 1 Summary of Precision Statements for Retroreflectance Data 1

Statistic Standard Deviation

(mcd/m2/lux) Coefficient of Variation (%)

d2s (mcd/m2/lux)

d2s% (%)

Values Within Between Within Between Within Between Within Between

12.0 18.8 2.8 4.7 33.9 53.3 7.8 13.3 2 Based on the results provided above, bias and precision statements for both the handheld 3 retroreflectometers and MRU are established in the following section. 4 5 6 BIAS AND PRECISION STATEMENT FOR THE MRU 7 8 In accordance with the methodology described in ASTM C-670 (15), the results from FDOT MRUs were 9 compared with reference values from handheld retroreflectometers and no statistically significant bias 10 was found. 11 12

The overall pooled results of two properly conducted retroreflectivity tests using the same MRU 13 on the same pavement marking test section should not differ by more than 7.8% (34.0 mcd/m2/lux for the 14 ten sites tested for this study) at a 95 percent confidence level for retroreflectivity values ranging between 15 200 and 800 mcd/m2/lux. 16

17 The overall pooled results of two properly conducted retroreflectivity tests using different MRUs 18

on the same pavement marking test section should not differ by more than 13.3% (53.0 mcd/m2/lux for 19 the ten sites tested for this study) at a 95 percent confidence level for retroreflectivity values ranging 20 between 200 and 800 mcd/m2/lux. 21 22 23 CONCLUSIONS 24 25 The present study was aimed at establishing the precision statement of the Mobile Retroreflectometer 26 Unit (MRU) for pavement marking retroreflectivity. For the precision and bias of the MRU, ten test 27 sections were selected to perform retroreflective measurements using the average results of three 28 handheld retroreflectometers as reference, in accordance with ASTM E-1710 (12). The results of two 29 MRUs were used as the basis for evaluating the repeatability and reproducibility of the MRU. 30 31

The average pavement marking retroreflectivity for the test sections ranged from 200 to 800 32 mcd/m2/lux. The MRU showed no statistically significant bias indicating that the device produces similar 33 retroreflectance measurements as the handheld retroreflectometer. The overall pooled standard deviation 34 for the retroreflectivity for the MRU was determined to be 12.0 mcd/m2/lux (within MRU) and 18.8 35 mcd/m2/lux (between MRUs). Also, the overall pooled coefficient of variance for the retroreflectivity for 36 the MRU was determined to be 2.8% (within MRU) and 4.7% (between MRUs). Therefore, the results of 37 two properly conducted retroreflectivity tests using the same MRU on the same pavement marking test 38 section should not differ by more than 7.8 percent at a 95 percent confidence level when using the same 39 MRU. In addition, the results of two properly conducted retroreflectivity tests using two different MRUs 40 on the same pavement marking test section should not differ by more than 13.3 percent at a 95 percent 41 confidence level. 42 43 44

45

TRB 2013 Annual Meeting Paper revised from original submittal.

Choubane, Sevearance, Lee, Upshaw, and Fletcher 14

ACKNOWLEDGMENTS 1 2 The work represented herein was the result of a team effort. The authors would like to acknowledge 3 Phillip Armand, Clayton Riddick, Adam Franklin, Kyle Kroodsma, James Greene, and Charles 4 Holzschuher for their assistance with the data collection effort and technical advice. 5 6 7 DISCLAIMER 8 9 The content of this paper reflects the views of the authors who are solely responsible for the facts and 10 accuracy of the data as well as for the opinions, findings and conclusions presented herein. The contents 11 do not necessarily reflect the official views or policies of the Florida Department of Transportation. This 12 paper does not constitute a standard, specification, or regulation. In addition, the above listed agency 13 assumes no liability for its contents or use thereof. 14 15

16

TRB 2013 Annual Meeting Paper revised from original submittal.

Choubane, Sevearance, Lee, Upshaw, and Fletcher 15

REFERENCES 1 2

1. Federal Highway Administration. FHWA Docket No. FHWS-2009-0139 (23 CFR Part 655 in 3 Federal Register Vol. 75, No. 77, dated April, 2010) 4

2. FHWA. Proposed Pavement Marking Retroreflectivity MUTCD Text. Revision 1. 5 http://mutcd.fhwa.dot.gov/knowledge/proposed09mutcdrev1/propmretromutcdxtxt.pdf. Accessed 6 June 28, 2012. 7

3. Holzschuher, C., and T. Simmons. Mobile Retroreflectivity Characteristics for Pavement 8 Marking at Highway Speeds, Florida Department of Transportation, 2005, pp. 6-14. 9

4. Pike, A. Evaluating Factors that may Influence the Accuracy of Mobile Retroreflectivity Data 10 Collection. Transportation Research Board. CD-ROM. No. 09-0493, Transportation Research 11 Board of the National Academies, Washington, D.C., 2009. 12

5. Fletcher, J., A. Philpott, B. Choubane, and C. Holzschuher. Characterization and Mitigation of 13 Temperature Sensitivity within a Mobile Retroreflectometer Unit. In Transportation Research 14 Record: Journal of Transportation Research Board, No. 2015, Transportation Research Board of 15 the National Academies, Washington, D.C., 2007, pp. 91-102. 16

6. Fletcher, J., Sevearance, J., Choubane, B., and C. Holzschuher. Characteristics of a Calibration 17 Standard for the Mobile Retroreflectometer Unit. Transportation Research Board. CD-ROM. No. 18 09-2103, Transportation Research Board of the National Academies, Washington, D.C., 2009. 19

7. Fletcher, J., Gonos, J., Choubane, B., and C. Holzschuher. Calibration Standards for Mobile 20 Retroreflectometer Unit. Transportation Research Board. CD-ROM. No. 08-2779, 21 Transportation Research Board of the National Academies, Washington, D.C., 2008. 22

8. Holzschuher, C., and A. Philpott. Mobile Retroreflectivity Unit Surveying of Maintenance Rating 23 Program Sites, Florida Department of Transportation, 2006, pp. 13. 24

9. Technical Evaluation Report; Summary of Evaluation Findings for 30-Meter Handheld and 25 Mobile Pavement Marking Retroreflectometers, HITEC Report #40525, March 2001. 26

10. Holzschuher, C., B. Choubane, J. Fletcher, J. Sevearance, and H. Lee. Repeatability of Mobile 27 Retroreflectometer Unit for Measurements of Pavement Markings. In Transportation Research 28 Record: Journal of Transportation Research Board, No. 2169, Transportation Research Board of 29 the National Academies, Washington, D.C., 2010, pp. 95-106. 30

11. Holzschuher, C., and T. Simmons. Mobile Retroreflectivity Characteristics for Pavement 31 Marking at Highway Speeds. Research Report FL/DOT/SMO/05-486, Florida Department of 32 Transportation, 2005. 33

12. ASTM E 1710-11. Standard Test Method for Measurement of Retroreflective Pavement Marking 34 Materials with CEN-Prescribed Geometry Using Portable Retroreflectometer. American Society 35 for Testing and Materials, ASTM International, West Conshohocken, Pa. 36

13. Gamma Scientific Laserlux 6 Users Guide. Gamma Scientific. San Diego, CA, 2002. 37 14. ASTM E 177-10. Standard Practice for Use of the Terms Precision and Bias in ASTM Test 38

Methods. American Society for Testing and Materials, ASTM International, West Conshohocken, 39 Pa. 40

15. ASTM C 670-10. Standard Practice for Preparing Precision and Bias Statements for Test 41 Methods for Construction Materials. American Society for Testing and Materials, ASTM 42 International, West Conshohocken, Pa. 43

16. ASTM C 802-09a. Standard Practice for Conducting an Interlaboratory Test Program to 44 Determine the Precision of Test Methods for Construction Materials. American Society for 45 Testing and Materials, ASTM International, West Conshohocken, Pa. 46

TRB 2013 Annual Meeting Paper revised from original submittal.