Automated Data Collection and Quality Management for Pavement Condition Reporting (2022)

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.

5   This chapter presents the findings of the literature review relevant to transitioning from man- ual to automated pavement condition surveys, agency DQMPs, and national reporting require- ments for pavement condition. Assessment of Pavement Condition There are essentially two broad categories for assessing pavement condition: manual and auto- mated surveys. Manual surveys are conducted by walking, cycling, or driving at slow speeds and noting the surface distress type, severity, and extent. Commonly used manual methods for pavement condition surveys include ASTM D6433, Standard Practice for Roads and Parking Lots Pavement Condition Index Surveys (ASTM 2020), Distress Identification Manual for the Long-Term Pavement Performance Program (Miller and Bellinger 2014), as well as agency- specified criteria. APCSs are conducted with specially designed vehicles to capture pavement profile data and surface images for a single lane in a single pass at posted speeds. The APCS profile data [e.g., rutting, faulting, international roughness index (IRI)] are typically processed in real time during data collection. Pavement images for quantifying surface distress (e.g., cracking, patching) are processed by using either semiautomated or fully automated methods. AASHTO and ASTM standards related to APCSs include the following: • AASHTO M 328, Standard Equipment Specification for Inertial Profiler (2018a), • AASHTO R 36, Standard Practice for Evaluating Faulting of Concrete Pavements (2021a), • AASHTO R 43, Standard Practice for Quantifying Roughness of Pavements (2021b), • AASHTO R 48, Standard Practice for Determining Rut Depth in Pavements (2013a), • AASHTO R 55, Standard Practice for Quantifying Cracks in Asphalt Pavement Surface (2013b), • AASHTO R 56, Standard Practice for Certification of Inertial Profiling System (2018b), • AASHTO R 57, Standard Practice for Operating Inertial Profiling Systems (2018f), • AASHTO R 85, Standard Practice for Quantifying Cracks in Asphalt Pavement Surfaces from Collected Pavement Images Utilizing Automated Methods (2018g), • AASHTO R 86, Standard Practice for Collecting Images of Pavement Surfaces for Distress Detection (2018c), • AASHTO R 87, Standard Practice for Determining Pavement Deformation Parameters and Cross Slope from Collected Transverse Profiles (2018e), • AASHTO R 88, Standard Practice for Collecting the Transverse Pavement Profile (2018d), • ASTM E950, Standard Test Method for Measuring the Longitudinal Profile of Traveled Surfaces with an Accelerometer-Established Inertial Profiling Reference (2018), C H A P T E R   2 Literature Review

6 Automated Data Collection and Quality Management for Pavement Condition Reporting • ASTM E1489, Standard Practice for Computing Ride Number of Roads from Longitudinal Profile Measurements Made by an Inertial Profile Measuring Device (2019), • ASTM E1656, Standard Guide for Classification of Automated Pavement Condition Survey Equipment (2016), • ASTM E1845, Standard Practice for Calculating Pavement Macrotexture Mean Profile Depth (2015), • ASTM E1926, Standard Practice for Computing International Roughness Index of Roads from Longitudinal Profile Measurements (2021a), • ASTM E3303, Standard Practice for Generating Pavement Surface Cracking Indices from Digital Images (2021b), • ASTM WK66179, New Specification for Generating Pavement Condition Indices from Digital Images (n.d.). Semiautomated methods include trained personnel using computer software and hardware to view the downward images and note pavement distress type and severity. Typically, the pave- ment rater notes the beginning and ending points of any observed distress, and the computer software calculates the extent of the distress. Fully automated methods require little to no human inter action for distress identification. Fully automated methods use both 2D and 3D technology. Fully automated systems typically include 2D systems to capture the intensity of reflected light (used to identify paint stripes, cracks, aggregate, etc.) and 3D systems to capture elevation or range (used to identify cracks, spalls, potholes, etc.). An example image from a fully automated APCS is shown in Figure 1. a. 2D b. 3D c. Identified distress Figure 1. Example 2D/3D image (Pierce and Weitzel 2019).

Literature Review 7   Transitioning from Manual to Automated Pavement Condition Surveys Timm and McQueen (2004) conducted an agency survey to determine the methods used by other SHAs for conducting pavement condition surveys. At that time, several agencies indicated they were in the process of transitioning or had already transitioned from manual to automated pavement surveys; however, APCSs were a relatively new method and had yet to be standard- ized. In total, 27 SHAs responded to an agency questionnaire related to the data collected as part of the pavement condition survey. Of these agencies, the majority (22 agencies) collected data on IRI and rut depth, slightly more than half (15 agencies) assessed surface cracking, and less than half (11 agencies) assessed faulting; two agencies assessed friction and two reported using APCS to collect other data. While many SHAs were optimistic that the transition to an APCS would improve the safety and efficiency of data collection, many SHAs noted potential implementation roadblocks. SHAs noted the two main issues for transitioning to an APCS included the lack of method standardization and the lack of information related to agencies that had successfully made the transition to automated surveys. Sivaneswaran and colleagues (2004) conducted an analysis for the transition from manual surveying to APCS for the Washington State DOT. As with the Timm and McQueen (2004) findings, the Washington State DOT noted a number of challenges in transitioning to an APCS, including compatibility with the historical windshield survey data and performance modeling, the cost and time frame for data collection and analysis, and quality control procedures. At the time of this analysis, the Washington State DOT had conducted 5 years of data collection with the APCS. A summary of findings included the following: • Cost and time comparison. The manual windshield survey was conducted by four teams of two, which required approximately 3,820 person-hours per year, along with 300 person-hours for data entry. The associated cost for the windshield survey was about $230,000 per year, plus about $38,000 per year for profile data collection, or about $268,000 per year. In comparison, the APCS required approximately 660 person-hours to collect the data and 2,600 person- hours to conduct the semiautomated analysis, with a total cost of approximately $240,000 (purchase cost of data collection vehicle excluded). • Quality control. Quality control for the windshield survey proved to be very difficult, since it required re-driving the section to confirm distress type, severity, and extent. However, re-driving an adequate sample size, estimated to be about 3.6%, would require an additional 150 person-hours to complete the data collection plus significantly more hours to travel to randomly sampled locations across the state. Due to this challenge, systemwide quality con- trol checks were rarely conducted. With the implementation of the APCS, quality control was dramatically improved as compared with the windshield survey, and the DOT was able to implement quality control checks on a random sample of approximately 6%. Figure 2 shows a comparison of production and sample pavement structural condition (PSC). Of the 459 samples, 434 (95%) had a PSC difference of less than 10 points (on a scale from 100 to 0). In Figure 2, the solid line represents the line of equality (R2 = 90.3%), and the dashed lines represent ±10 PSC points. • Compatibility with historical data. The DOT was concerned with potential differences due to the effect of viewing through the windshield at slow speeds versus playing images back on computer workstations and noting the distress type and severity. Assessment of the effect of using preselected extent ranges and the predominant distress for the windshield survey compared with the exact extent determined with the APCS was also of concern. Figure 3 illustrates the cumulative lane kilometers of asphalt pavement as a function of PSC for the 1998 windshield survey and the 1999 APCS. Asphalt pavements accounted for more than 60% of the total state highway network and accounted for about 70% of the total preservation

8 Automated Data Collection and Quality Management for Pavement Condition Reporting funding. The network PSC from the 1998 windshield survey and the 1999 APCS agreed rea- sonably well, and in particular for PSC values of less than 60 (the threshold for triggering rehabilitation). The differences at higher PSC values were attributed to the preselected extent ranges from the windshield survey, which generally overestimated the amount of distress or underestimated the PSC. However, the differences at higher PSC values were still within 5 to 10 PSC points. Chan and colleagues (2016) noted the Ontario Ministry of Transportation initiated the tran- sition from manual to APCS beginning in 2013 (full implementation in 2015). The Ministry conducted a comparison and validation analysis of specific distress from the manual and APCS data sets. In total, 934 pavement segments were surveyed by using both manual survey methods and APCS. In order to provide compatible distress ratings for the pavement management system, the APCS distress rating [i.e., pavement condition index (PCI), distress manifestation index (subject rating index from 10 to 0)] was converted to an equivalent manual distress rating. The study found that using only the APCS cracking results (11 types of crack distress) provided an acceptable distress rating for network-level pavement management activities. 20x103 15 10 5 0 C um ul at iv e La ne K ilo m et er s 100806040200 PSC Survey 1998 - Windshield Survey 1999 - Automated Figure 3. Cumulative lane kilometers of asphalt pavement as a function of PSC (Sivaneswaran et al. 2004). R2 = 90.3% 0 20 40 60 80 100 0 20 40 60 80 100 PSC Sample PS C Pr od uc tio n Figure 2. Comparison of sample and production PSC (Sivaneswaran et al. 2004).

Literature Review 9   Vavrik et al. (2013) investigated the feasibility of transitioning from manual to semiautomated pavement condition surveys and analysis for the Ohio DOT. Numerical and statistical com- parisons of results from DOT raters and three data collection vendors were conducted to quan- tify pavement distress, severity, and extent on 44 test sites in Ohio. Vendors matched the Ohio DOT–rated distresses 74.5% of the time, matched distress type and severity 33.4% of the time, and matched distress type, severity, and extent only 13.5% of the time. The study indicated the match between data collection vendors and the Ohio DOT raters was insufficient for a direct transition from manual to semiautomated condition surveys. To make the transition, modifica- tions would need to be made to the Ohio DOT’s pavement condition rating (PCR) index, deci- sion trees, performance models, and pavement management software. While improvements in pavement data collection were identified as being needed, several benefits of transitioning to semiautomated pavement condition surveys were noted: • Increased rater safety; • Improved data accuracy for certain distresses, severities, and extents (e.g., rutting and faulting); • Enhanced timeliness of data collection and processing; • Ability to easily track, review, and reprocess historical data and images; • Ability to collect data compatible with AASHTO requirements for Pavement ME Design; • Ability to collect data compatible with Highway Performance Monitoring System (HPMS) requirements; • District access to pavement images for project-level reviews; • Consistent, well-defined methods for future automated identification of distress, severity, and extent; • District access to vendors for ancillary data collection; and • Ability to combine IRI, rutting, and asset collection with pavement distress ratings. Pierce and Weitzel (2019) conducted a survey of highway agencies in the United States, Canada, and Puerto Rico to capture experience with APCSs. As of 2018, 16 agencies used fully automated survey methods, 21 used a combination of fully and semiautomated methods, 6 used a combination of manual and automated methods, and 6 used manual survey methods only. Of the responding agencies, slightly more than half indicated using fully automated methods for asphalt pavements, and slightly less than half reported using fully automated methods for jointed plain concrete pavement (JPCP) and continuously reinforced concrete pavement (CRCP) (Table 1). Furthermore, agencies were asked to identify which survey method was used for each distress type for each pavement type. The survey results are shown in Table 2, Table 3, and Table 4 for asphalt, JPCP, and CRCP, respectively. For asphalt pavements, most of the agencies indicated they used either fully or semiautomated methods for all distress types, except for corrugation, weathering, and polished aggregate, for which manual survey methods were more commonly used. For JPCP, semi- and fully automated methods were used to evaluate the majority of distress Pavement Type Number of Agencies Fully Automated Fully & Semiautomated Total Responses Asphalt 26 24 50 JPCP 18 23 41 CRCP 8 11 19 Table 1. Summary of automated method used, by surface type: 2018.

10 Automated Data Collection and Quality Management for Pavement Condition Reporting types, except for pumping, scaling, and shrinkage cracks. For CRCP, semi- and fully automated survey methods were predominantly used to assess distress type, severity, and extent. The purpose of the 2019 synthesis was not only to evaluate the state of practice, but also to capture the agencies’ experiences with APCS. Agencies identified several benefits of APCS, including improved safety, efficiency, and consistency; satisfaction with identification of crack type; and the ability for data and images to be used by various agency users for multiple applica- tions. The surveyed agencies also found APCS to be a useful tool in identifying projects for state transportation improvement programs. However, challenges with using APCS were also identified. These included difficulty in deter- mining reasonable data quality tolerances, establishing reference measurements, consistently measuring rut depth, configuring protocols and distress detection algorithms for the new data sets and performance criteria, producing meaningful reports and project assessments that also met the report requirements, and maintaining consistency from year to year and from vendor to vendor. National Reporting Requirements SHAs are required to report pavement condition by several national programs, including FHWA under PM2 and the HPMS and the Governmental Accounting Standards Board (GASB). Reporting requirements for each program are briefly described in the following sections. Measure Fully Automated Semiautomated Manual Total No. of Responses 35 gnittuR 0 3 56 55 IRI 0 0 55 Transverse cracking 32 13 10 55 Alligator cracking 29 15 10 54 Longitudinal cracking 33 9 9 51 41 selohtoP 13 9 36 01 gnihctaP 15 11 36 41 gnilevaR 11 10 35 Block cracking 16 11 7 34 Edge cracking 19 10 4 33 Cross slope 30 0 1 31 01 gnideelB 9 9 28 Reflection cracking 16 7 4 27 91 erutxeT 1 2 22 Lane/shoulder drop-off 9 3 5 17 8 noisserpeD 2 3 13 5 gnivohS 2 6 13 Bumps and sags 8 1 2 11 3 noitagurroC 2 6 11 0 gnirehtaeW 3 7 10 Polished aggregate 1 3 4 8 Faulting (composite pavements) 4 0 0 4 2 noitanimaleD 0 0 2 Wheel path cracking 1 1 0 2 Source: Pierce and Weitzel (2019). Table 2. Method used for asphalt pavements.

Literature Review 11   Measure Fully Automated Semiautomated Manual Total No. of Responses 91 IRI 0 0 19 Cross slope 9 0 0 9 Longitudinal cracking 8 7 2 17 Transverse cracking 6 6 1 13 Texture 6 0 1 7 5 tuohcnuP 8 1 14 Lane/shoulder 8 2 1 5 ffo-pord Spalling 3 4 1 8 3 gnihctaP 7 2 12 Durability 3 3 2 8 Scaling 1 1 1 3 Map cracking 1 3 0 4 Polished aggregate 0 2 1 3 Blowups 0 4 2 6 Source: Pierce and Weitzel (2019). Table 4. Method used for CRCP. Measure Fully Automated Semiautomated Manual Total No. of Responses 44 IRI 0 0 44 73 gnitluaF 3 2 42 Cross slope 20 1 1 22 Longitudinal cracking 20 13 7 40 Transverse cracking 16 17 6 39 21 erutxeT 1 2 15 8 gnihctaP 14 7 29 Corner cracking 7 16 7 30 7 gnillapS 15 8 30 Joint seal damage 6 7 7 20 Lane/shoulder drop-off 6 4 5 15 4 ytilibaruD 9 6 19 Map cracking 4 7 2 13 2 spuwolB 6 3 11 2 gnipmuP 3 6 11 Broken slabs or percent cracked slabs 1 3 0 4 Polished aggregate 1 3 3 7 1 gnilacS 3 7 11 Shattered area or slabs 1 2 0 3 Shrinkage cracks 0 0 2 2 Source: Pierce and Weitzel (2019). Table 3. Method used for JPCP.

12 Automated Data Collection and Quality Management for Pavement Condition Reporting National Performance Management Measures Enacted in 2018, PM2 established a performance- and outcome-based program that requires SHAs and metropolitan planning organizations (MPOs) to prepare and use federally established performance measures. Specific to pavements, PM2 requires agencies to submit pavement con- dition on the Interstate highway system and the non-Interstate NHS to FHWA, in accordance with the HPMS Field Manual (FHWA 2018b). Performance Measures Required pavement performance measures for asphalt pavements, JPCP, and CRCP are sum- marized in Table 5 through Table 7. For asphalt pavements, performance measures include the IRI, rut depth, and percent cracking. (Performance measures are further defined in the HPMS section of this chapter.) For JPCP, the performance measures include IRI, faulting, and percent Table 5. Asphalt pavement performance measures. Measure Unit Test Procedure/Process Good Fair Poor IRI in./mi • AASHTO M 328 (2018a) and • AASHTO R 57 (2018f) <95 Rutting in. • AASHTO R 48 (2013a) or • AASHTO PP 70 (2017a)a <0.2 Cracking Percent area • Methodsb and Protocolsc • Fatigue cracking in wheel paths  <5 >170 >0.4 >20 aReplaced by AASHTO R 88 (2018d). bWindshield survey, visual distress survey, manually identified by video, automated crack identification to detect cracking from video, combined manual and automatic crack identification from video, and/or 3D imaging system. cLong-Term Pavement Performance (LTPP), AASHTO, modified LTPP, modified AASHTO, and/or state-developed protocol. 95–170 0.2–0.4 5–20 Table 7. Continuously reinforced concrete pavement performance measures. >170 >10 95–170 5–10 Measure Unit Test Procedure/Process Good Fair Poor IRI in./mi • AASHTO M 328 (2018a) and • AASHTO R 57 (2018f) <95 Cracking Percent area • Methodsa and Protocolsb • Visible distressc <5 aWindshield survey, visual distress survey, manually identified by video, automated crack identification to detect cracking from video, combined manual and automatic crack identification from video, and/or 3D imaging system. bLTPP, AASHTO, Modified LTPP, Modified AASHTO, and/or state-developed protocol. cIncludes only longitudinal cracking, punchouts, spalling, or other visible defects. >170 >0.15 >15 95–170 0.10–0.15 5–15 Measure Unit Test Procedure/Process Good Fair Poor IRI in./mi • AASHTO M 328 (2018a) and • AASHTO R 57 (2018f) <95 Faulting in. • AASHTO R 36 (2021a) <0.1 Cracking Percent slabs • Methodsa and Protocolsb • Transverse cracked slabsc <5 aWindshield survey, visual distress survey, manually identified by video, automated crack identification to detect cracking from video, combined manual and automatic crack identification from video, and/or 3D imaging system. bLTPP, AASHTO, Modified LTPP, Modified AASHTO, and/or state-developed protocol. cIncludes partial slabs with cracking over majority of slab length. Table 6. Jointed plain concrete pavement performance measures.

Literature Review 13   cracking. For CRCP, the performance measures are IRI and percent cracking. Each performance measure is defined according to condition: good, fair, or poor. For the Interstate highway system and non-Interstate NHS with posted speeds of less than 40 miles per hour (mph), agencies have the option to report pavement condition according to the present serviceability rating (PSR), in accordance with HPMS requirements, in lieu of IRI, rutting, faulting, and cracking. PSR performance measures are defined as follows: • Good: PSR > 4.0, • Fair: > 2.0 PSR < 4.0, and • Poor: PSR ≤ 2.0. PSR is further discussed in the HPMS section of this chapter. Data Collection Requirements Table 8 provides a summary of data collection requirements for both Interstate and non- Interstate routes in the NHS. Data collection requirements are essentially the same for Inter- state and non-Interstate NHS routes, except for the direction of travel and the frequency of data collection (shown in bold italicized text in Table 8). For Interstate routes, data collection is to occur annually in at least one direction of travel. SHAs may also collect and report pavement condition for both directions of divided highways on the Interstate system. For non-Interstate routes, data collection is conducted biennially and in one direction of travel. Overall Pavement Condition Determination of overall pavement condition is based on the predominant number of perfor- mance measures in a given condition category (Table 9). Figure 4 provides an example for determining overlay pavement condition. In this example, an asphalt pavement segment has an average IRI value of 180 in./mi (poor condition), 7% crack- ing (fair condition), and average rutting of 0.3 inches (fair condition). Therefore, since two of the three performance measures are in fair condition, the overall pavement condition would be fair. As stated in 23 CFR § 490.315, the percent of pavement lane miles, on the Interstate system, in poor condition, “shall not exceed 5.0 percent” and “shall not exceed 10.0 percent” for the state of Alaska. etatsretnI-noN etatsretnI NHS Routes • Full extent of mainline highway • Rightmost or continuous through lane • Continuous data collection • At least one direction of travel • Annual frequency • Sampling is not allowed • Averaging across directions is not allowed • Reported in 0.10-mi segments • Excludes bridges and other pavement types (e.g., brick, gravel) • Follow HPMS Field Manual (FHWA 2018b) • IRI, rut depth, faulting, percent cracking • Surface and structure type • No more than 5% missing, invalid, or unresolved data • Full extent of mainline highway • Rightmost or continuous through lane • Continuous data collection • One direction of travel • Biennial frequency • Sampling is not allowed • Averaging across directions is not allowed • Reported in 0.10-mi segments • Excludes bridges and other pavement types (e.g., brick, gravel) • Follow HPMS Field Manual (FHWA 2018b) • IRI, rut depth, faulting, percent cracking • Surface and structure type • No more than 5% missing, invalid, or unresolved data Table 8. Comparison of Interstate and non-Interstate data collection requirements.

14 Automated Data Collection and Quality Management for Pavement Condition Reporting Reporting Requirements In accordance with 23 CFR § 490.319, each SHA is required to report to FHWA, no later than April 15 of each year, the information necessary to calculate the Interstate system condition measures and, no later than June 15 of each year, the information on the non-Interstate NHS condition measures. In addition, MPOs are to report targets to their respective SHAs and to report baseline pavement condition and progress toward achieving the targets. Table 10 provides a summary of agency reporting requirements, and Figure 5 provides a sum- mary of pavement condition reporting requirements. Highway Performance Monitoring System FHWA is responsible for providing highway transportation data and performance infor- mation in support of apportioning Federal-Aid Highway Program funds. In support of this requirement, the HPMS was adopted in 1978 as a national highway transportation program (FHWA 2018b). Data contained within the HPMS include all federal, state, county, city, and privately owned (e.g., toll) roads; however, more detailed information is required for the Inter- state system and non-Interstate NHS. SHAs are required to submit HPMS data no later than June 1 of each year to FHWA. Specific to pavements, the HPMS data are used to determine the Condition Asphalt JPCP CRCP Good Good ratings for all three conditionsa; PSRb ≥ 4.0 Good ratings for all three conditionsc; PSR ≥ 4.0 Good ratings for both conditionsd; PSR ≥ 4.0 Fair Good or poor conditions are not met; PSR > 2.0 and < 4.0 Good or poor conditions are not met; PSR > 2.0 and < 4.0 Good or poor conditions are not met; PSR > 2.0 and < 4.0 Poor Two or more ratings are in poor condition; PSR ≤ 2.0 Two or more ratings are in poor condition; PSR ≤ 2.0 Poor ratings for both conditions; PSR ≤ 2.0 aThe three conditions evaluated for asphalt are IRI, rutting, and cracking. bIn accordance with the HPMS Field Manual (FHWA 2018b). cThe three conditions evaluated for JPCP are IRI, faulting, and cracking. dThe two conditions evaluated for CRCP are IRI and cracking. Table 9. Criteria for overall pavement condition. Cracking = 7.0% G: <5%; F: 5-20%; P: >20% Rutting = 0.3 in.IRI = 180 in./mile G: <95; F: 95-170; P: >170 G: <0.2"; F: 0.2"- 0.4"; P: >0.4" Figure 4. Example overall condition (Rodriguez et al. 2017).

Literature Review 15   percentage of pavements in good and poor condition on the Interstate system and the non- Interstate NHS. Pavement condition is quantified according to the following measures: • Average IRI: Average IRI is required for all pavement types. IRI data collection is conducted in accordance with AASHTO M 328 (2018a), AASHTO R 56 (2018b), and AASHTO R 57 (2018f) and reported in accordance with AASHTO R 43 (2021b). • Asphalt Pavement Rut Depth: Average rut depth is collected in accordance with AASHTO R 48, by using no less than five profile points, or in accordance with AASHTO PP 69 and AASHTO PP 70 (2013a, 2017b, 2017a). The maximum spacing between consecutive trans- verse profile measurements is ≤12 inches. Rut depths are calculated for both wheel paths and reported as the average of the two wheel paths. Figure 6 illustrates the rut depth measurement. • JPCP Faulting: Fault measurements are in accordance with AASHTO R 36, avoiding mea- surement of faulting at cracks (Figure 7) (AASHTO 2021a). Fault height is calculated using LTPP 25-mm-interval profiler data (Method A) or 20.7-mm-interval high-speed inertial pro- filer data (Method B) and averaged over the right wheel path. • Percent Cracking: Percent cracking for asphalt pavements can be conducted with manual methods or APCS methods. However, APCS methods are preferred at locations where IRI measurements are conducted. Cracking is collected in accordance with AASHTO R 85 and Figure 5. PM2 pavement condition reporting requirements. Interstate System • Every year • Full extent • One lane • At least one direction Non-Interstate NHS • Every 2 years • Full extent • One lane • One direction Required Pavement Condition Data • IRI • Percent Cracking • Rutting • Faulting Table 10. PM2 reporting requirements. Baseline (2018) Midpoint (2020) Full (2022) • 2- and 4-year targets, and the basis for each target. • Baseline condition/performance from the latest data collection. • How the established targets support the TAMP. • Boundary of urbanized areas and latest Census population data. • Due Oct. 1, 2018, and Oct. 1 every 4 years. • 2-year actual condition data. • 2-year progress in achieving targets, actual vs. target, and why different. • Effectiveness of the investment strategies. • When applicable, submit adjusted 4-year target, basis for the adjustment, and how it supports the TAMP. • Progress toward achieving 2- year targets, prior accomplishments, and planned activities for 4-year targets. • Extenuating circumstance(s). • Due Oct. 1, 2020, and Oct. 1 every 4 years. • Actual 4-year condition/performance data. • Progress toward 4-year targets, actual vs. targets, and why different. • Effectiveness of the investment strategies in TAMP for the NHS. • Extenuating circumstances. • Due Oct. 1, 2022, and Oct. 1 every 4 years. Note: TAMP = Transportation Asset Management Plan.

16 Automated Data Collection and Quality Management for Pavement Condition Reporting Figure 7. JPCP faulting measurements (FHWA 2018b). Figure 6. Asphalt pavement rutting measurements (FHWA 2018b).

Literature Review 17   Figure 9. JPCP: percent slabs with transverse cracking (adapted from FHWA 2018b). Figure 8. Asphalt pavement: percent area wheel path fatigue cracking (FHWA 2018b). AASHTO R 86 (2018g, 2018c). Percent cracking is defined as the total area of fatigue cracking (alligator and longitudinal), in both wheel paths, for all severity levels and sealed and unsealed cracks (Figure 8). For JPCP, manual or APCS methods that can identify at least 85% of all slab cracking are used. Percent cracking is based on the number of slabs containing one or more transverse cracks, which extend for at least one-half the lane width, divided by the total number of slabs in the section. Longitudinal cracks, corner breaks, D-cracking, and cracking due to alkali silica reactivity are not included in the percent cracking calculation (Figure 9). Percent cracking for CRCP is determined by using manual or APCS methods that can iden- tify at least 85% of all surface distress. Percent cracking is based on the area of longitudinal cracking (length × 1-foot wide), and area of punchouts (area between two adjacent transverse cracks and the edge of the pavement or longitudinal joint) and patching (Figure 10 through Figure 12). • Present serviceability rating: For sections on the NHS where the posted speed limit is less than 40 mph, agencies may report PSR in lieu of IRI, rutting, faulting, and percent cracking. Table 11 provides a description of pavement distress and the associated PSR ranges. In 2021, FHWA prepared a report and developed a Microsoft® Excel workbook to support agencies (e.g., SHAs and MPOs) in quantifying pavement distress according to PSR (and PCI) definitions. Distress types used to calculate PSR (and PCI) for both asphalt pavements and JPCP are summarized in Table 12.

18 Automated Data Collection and Quality Management for Pavement Condition Reporting Figure 12. CRCP percent cracking: patching (FHWA 2018b). Figure 10. CRCP percent cracking: longitudinal cracking (adapted from FHWA 2018b). Figure 11. CRCP percent cracking: punchouts (FHWA 2018b).

Literature Review 19   The FHWA process for determining PSR and PCI from limited distress types is based on procedures developed by Mok and Smith (1997) and the San Francisco Bay Area Metropolitan Transportation Commission (MTC), respectively. The general equation for determining PCI under ASTM D6433 is shown in Equation 1 (ASTM 2020). In addition, under the MTC PCI method, only the distress types summarized in Table 12 are used in the PCI calculation. ∑∑ ( ) ( )= − ==PCI , , , (1)11C a T S D F t qi j ijjmip i where PCI = pavement condition index; C = maximum value of the condition index (no distress = 100); a(T, S, D) = deduct value function; varies on basis of distress type (T), severity (S), and den- sity (D); F(t,q) = adjustment function; varies with total deduct value (t) and number of deduct values (q); i, j = counters for distress types and severity levels, respectively; p = total number of observed distress types; and mi = number of severity levels for the ith distress type; typically includes low-, medium-, and high-severity distress. Asphalt Pavement Distress Types JPCP Distress Types Alligator cracking Corner break Block cracking Divided slab gnitluaF snoitrotsiD Longitudinal and transverse cracking Linear cracking Patching and utility cut patching Patching and utility cuts Rutting and depressions Scaling, map cracking, crazing gnillapS gnilevaR Weathering Source: Thyagarajan et al. (2021). Table 12. Distress types used to determine PSR and PCI. PSR Description 5.0–4.0 Only new (or nearly new) superior pavements are likely to be smooth enough and distress free (sufficiently free of cracks and patches) to qualify for this category. Most pavements constructed or resurfaced during the data year would normally be rated in this category. 4.0–3.0 Pavements in this category, although not quite as smooth as those described above, give a first class ride and exhibit few, if any, visible signs of surface deterioration. Flexible pavements may be beginning to show evidence of rutting and fine random cracks. Rigid pavements may be beginning to show evidence of slight surface deterioration, such as minor cracks and spalling. 3.0–2.0 The riding qualities of pavements in this category are noticeably inferior to those of new pavements and may be barely tolerable for high-speed traffic. Surface defects of flexible pavements may include rutting, map cracking, and extensive patching. Rigid pavements in this group may have a few joint failures, faulting and cracking, and some pumping. 2.0–1.0 Pavements in this category have deteriorated to affect the speed of free-flow traffic. Flexible pavement may have large potholes and deep cracks. Distress includes raveling, cracking, rutting, and occurs over 50% of the surface. Rigid pavement distress includes joint spalling, patching, cracking, scaling, and may include pumping and faulting. 1.0–0.1 Pavements in this category are in an extremely deteriorated condition. The facility is passable only at reduced speeds, and with considerable ride discomfort. Large potholes and deep cracks exist. Distress occurs over 75% or more of the surface. Source: FHWA (2018b). Table 11. PSR definition.

20 Automated Data Collection and Quality Management for Pavement Condition Reporting In the Mok and Smith (1997) procedure, a correlation between PSR and PCI was developed on the basis of the comparison of windshield survey results to determine PSR for pavement segments with PCI results based on the MTC method (e.g., ASTM D6433 with limited distress types). Developed correlation equations for asphalt and concrete pavements are shown in Equa- tions 2 and 3, respectively. PSR 0.55 ln 100 PCI 1 2.17 (2)adj( )= − × × −  + PSR 0.047 PCI 0.231 (3)= × + where PSR = present serviceability rating, PCI = pavement condition index (based on MTC limited distress types), and PCIadj = PCI calculation excluding low-severity cracking. In support of the conversion from limited distress types to PCI and PSR, FHWA developed a Microsoft® Excel® spreadsheet to assist agencies with the conversion process. A screenshot of the PSR estimation workbook is shown in Figure 13. Governmental Accounting Standards Board GASB is an “independent, private-sector organization that establishes accounting and finan- cial reporting standards for U.S. state and local governments” (GASB 2020). Since its origination in 1984, GASB has been working toward revising the model used by governments for finan- cial reporting. It was believed a new model was needed, as state and local government financial reports did not provide sufficient information regarding financial position and cost of services. To address the issue, a special task force, which included experts from AASHTO and FHWA, was assembled, and GASB-sponsored public hearings and focus groups were held. In June 1999, GASB issued Statement 34 (GASB-34), establishing a financial reporting stan- dard for state and local governments in the United States (GASB 1999). The required statements for the newly established reports include the following: • Management discussion and analysis, which introduces the financial statements, as well as an overview of the government agency’s financial activities. • Basic financial statements: – Government-wide financial statements, which include a statement of net assets and a state- ment of activities, and Units US Custom 60.6 Pavement Type AC 2.41 Sample Area in Sq.feet 1200 Table 1. Distress Type, Severity and Quantity Distress ID Distress Type Severity Quantity 1a Quantity 2a Quantity 3a Quantity 4a Quantity 5a Quantity 6a Quantity 7a 1 Alligator Cracking Low 100 0 0 0 0 0 0 10 Longitudinal & Transverse Cracking Medium 50 0 0 0 0 0 0 19 Raveling Medium 25 0 0 0 0 0 0 Pavement Condition Index (PCI) Present Serviceability Rating (PSR) Press Alt + c to clear All Data Clear Data Figure 13. Example PCI and PSR calculations (Thyagarajan et al. 2021).

Literature Review 21   – Fund financial statements, which include governmental funds, proprietary funds, and fidu- ciary funds. • Required supplementary information, which includes budgetary comparison schedules as well as any other information required in previous GASB statements. At a time when many government agencies were starting to deal with an aging infrastructure system while facing budgetary constraints, GASB-34 established the implementation of an asset management framework. This framework gave agencies the ability to make more strategic and cost-effective allocation decisions. Following the framework steps also assisted governments in meeting the requirements set forth in GASB-34. The basic steps of the generic asset management framework are as follows: 1. Performance expectations, which are consistent with goals, available budgets, and organiza- tion policies, are established. 2. Inventory and performance data are collected and analyzed. 3. Cost-effective strategies for budget allocation are produced. 4. Projects are selected and programs are implemented. Overall, GASB-34 was a significant step toward providing more accountability to the traveling public in relation to transportation infrastructure. It provided a framework for asset manage- ment as well as established new financial reporting standards. GASB-34 requires agencies to report general infrastructure assets along with depreciation or preservation costs (PBC et al. 2004). Fixed assets include nondepreciable land, land improve- ments, completed construction (90% or greater), in-progress nondepreciable construction, infrastructure, and licensed vehicles (GASB 2020). Agencies are required to report the following (GASB 2020): 1. Management discussion and analysis that introduces the basic financial statements and over- view of financial activities (comparisons of current year with previous year); 2. A basic financial statement consisting of assets, liabilities, revenues, expenses, and gains and losses; and 3. Required supplemental information, including budgetary comparison schedules and other data as required. Agencies have the option to estimate asset value using either straight line depreciation or the “modified approach,” which values assets by condition. For asset depreciation using the modified approach, agencies are required to have an asset management program that includes an up-to-date inventory of assets, a replicable condition assessment that uses a measurable scale, and annual costs for maintaining and preserving the assets at the agency-established con- dition level. An example GASB-34 agency report is shown in Figure 14. Agency Data Quality Management Plans As part of PM2, SHAs were also required to develop, submit (by May 20, 2018), and use a pavement condition DQMP, regardless of the pavement condition data collection method. Requirements of the DQMP included (at a minimum): • Data collection equipment calibration and certification; • Certification process for persons performing manual data collection; • Data quality control activities prior to and periodically during data collection;

22 Automated Data Collection and Quality Management for Pavement Condition Reporting • Data sampling, review, and checking processes; and • Error resolution procedures and data acceptance criteria. DQMPs for the SHAs, including Puerto Rico and the District of Columbia, were reviewed, and information regarding agency practices for assessing pavement condition was summarized in relation to • Types of pavement condition data collected, • Requirements for equipment calibration and certification, • Data quality control criteria, and • Data acceptance criteria. All agency DQMPs include IRI, rutting, and faulting (when applicable) in accordance with PM2 requirements. In addition, agency DQMPs include percent cracking and other surface distress types as well as other profile-based information (e.g., cross slope, macrotexture) in the DQMP. Table 13 and Table 14 summarize pavement surface distress, as reported in each SHA’s DQMP, for asphalt and concrete pavements, respectively. In addition to the PM2-required pavement condition measures, predominant distress types (≥20 agencies) collected for asphalt pavements include alligator cracking, longitudinal cracking, transverse cracking, and patching. In addition to IRI, faulting, and percent cracking, predominant distress types (≥20 agencies) collected for concrete pavements include longitudinal cracking, patching, spalling, and transverse cracking. Agency DQMPs also require inclusion of equipment calibration and certification, quality control, and acceptance criteria. Tables 15 through 22 provide a summary of equipment calibra- tion and certification, quality control, and acceptance requirements for IRI, rut depth, faulting, Functions Expenses Charges for Services Operating Grants and Contributions Capital Grants and Contributions Governmental Activities Business-Type Activities Total Component Units Governmental activities Function #1 $ $ $ $ -$ — -$ — Function #2 $ $ $ — -$ — -$ — Function #3 $ $ $ $ -$ — -$ — Total governmental activities $ $ $ $ -$ — -$ — Business-type activities BTA #1 $ $ — $ — $ $ — BTA #2 $ $ — $ — $ $ — Total business-type activities $ $ — $ — $ $ — Total primary government $ $ $ $ -$ $ $ — Component units CU #1 $ $ $ $ — — — $ $ $ $ $ $ — $ — $ — X — $ -$ — — $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ Transfers Change in net assets Net assets—beginning Net assets—ending Total general revenues, contributions, special items, and transfers Program Revenues Net (Expense) Revenues and Changes in Net Assets General revenues—detailed Contributions to permanent funds Special items Figure 14. Example GASB-34 statement of activities (adapted from GASB 1999).

Literature Review 23   DC DE FL GA HI IA ID IL IN KS LA MA MD ME MI MN MO MS MT NC ND NE NH NJ NM NV NY OH OK OR PA PR RI SC SD TN TX UT VA VT WA WI WV WY Total 50 50 33 34 9 18 7 3 33 8 26 10 17 6 35 Note: Crack. = cracking; Allig. = alligator; Long. = longitudinal; Refl. = reflection; Trans. = transverse. Source: Pierce and Weitzel (2019). State IRI Rut Percent Crack. Allig. Crack. Bleed- ing Block Crack. Edge Crack. Flush- ing Long. Crack. Long. Joint Crack. Patch- ing Pot- hole Ravel- ing Refl. Crack. Trans. Crack. AK AL AR AZ CA CT Table 13. Pavement distress data collected: asphalt-surfaced pavements.

24 Automated Data Collection and Quality Management for Pavement Condition Reporting State IRI Fault Percent Crack. Blowups/ Repair Corner Break Cracked Slabs Multi- cracked Slabs D- Crack. Delam- ination Diagonal Crack. Edge Distress Joint Distress AL AR AZ CA CT DC DE FL GA HI IA ID IL IN KS LA MA MD MI MN MO MS MT NC ND NE NJ NM NV NY OH OK OR PA PR RI SC SD TN TX UT VA WA WI WV WY Total 46 46 32 4 14 7 10 6 1 2 2 2 Source: Pierce and Weitzel (2019). Table 14. Pavement distress data collected: concrete-surfaced pavements.

Literature Review 25   and cracking. It should be noted the information contained in these tables does not necessarily represent all data quality elements required by each SHA. For example, the information on equipment calibration and certification is limited to IRI, rut depth, faulting, and cracking (i.e., percent cracking, asphalt pavement alligator cracking, concrete pavement transverse cracked slabs) and excludes summaries of other distress types (e.g., transverse cracking, longitudinal cracking, raveling, spalling), profile measures (e.g., macrotexture, cross slope), GPS, and image requirements. Similarly, quality control and acceptance (e.g., data completeness checks, expected value checks, correct distress by pavement type) are excluded from the tables for brevity. Finally, exact descriptions of data quality requirements have been shortened or abbreviated to fit within the summary tables. Although requirements for equipment calibration and certification, quality control, and accep- tance vary across all SHAs, the most commonly used criteria are summarized in Table 23. State Joint Seal Damage Long. Crack Map Crack Patch -ing Pop outs Pothole Pump -ing Punch- out Scal -ing Ln.-Shld. Drop-off Slab Replc. Spall- ing Trans. Crack AL AR AZ CA DE HI IA IL IN LA MA MD MI MN MS MT NC ND NE NJ NM NV OK OR PA SC TN TX UT WA WI WV Total 6 23 4 22 1 1 2 13 3 2 2 23 23 Note: Ln.-Shld. = lane shoulder; Replc. = replacement. Source: Pierce and Weitzel (2019). Table 14. (Continued).

26 Automated Data Collection and Quality Management for Pavement Condition Reporting Agencya Resolution Accuracy (to reference value) Repeatability AK (2018) 0.0004 in. ≥ )snur 01( %5 < DS %59 AL (2018) ns Cross-correlation > 88% Avg. 5 runs becomes target IRI AR (2018) 1 in./mi ±10% SD ± 5% (3 runs) AZ (2018) 0.0001 in. Contractor provided Contractor provided CA (2018) 1 in./mi ±10% ±5% (3 runs) CT (2018) 1 in./mi ±8% Mean ± 5% (5 runs) DC (2019) 1 in./mi ±5% Mean ± 5% DE (2018) 1 in./mi ±10 in./mi SD < 5 (10 runs) FL (2018) 1 in./mi Cross-correlation ≥ 90% or ±5% IRI Cross-correlation ≥ 0.92 GA (2018) 1 in./mi ±5% vendor baseline value ±5% (3 runs) IA (2018) 1 in./mi ±10% ±5% (3 runs) ID (2018) 1 in./mi ±5% ±5% (3 runs) IL (2018) Vendor certified Vendor certified Vendor certified IN (2019) 1 in./mi ±5% ±5% (3 runs) MA (2018) 1 in./mi ≥80% ≥90% (10 runs) MD (2018) ns Based on 40 sections Based on 40 sections ME (2018) 1 in./mi na Mean ± 10% (5 runs) MI (2018) ns Avg. ±5 in./mi (10 runs) SD ± 2 in./mi (10 runs) MO (2018) 1 in./mi ±5% ±5% (3 runs) MS (2019) 1 in./mi Cross-correlation ≥ 90% or ± 5% Cross-correlation ≥ 0.92 MT (2018) 1 in./mi ±5% ±5% (3 runs) NC (2018) 1 in./mi Bi %5< :noisicerP %5< :sa ND (2018) 1 in./mi ±5% ±5% (3 runs) NE (2018) 1 in./mi ±5% ±5% (3 runs) NH (2019) 1 in./mi ns ns NJ (2019) 1 in./mi ±5% ±3% (3 runs) NM (2018) 0.6 in./mi ≥ lacirotsih %01 < DS ,)snur 01( %5 < DS %09 NV (2018) 1 in./mi ±5% ns NY (2018) 1 in./mi ns ns OH (2018) ns ns Avg. ± 5% (5 runs) OK (2018) 1 in./mi ±5% ±5% (3 runs) OR (2018) 1 in./mi Cross-correlation ≥ 90% (5 runs) Cross-correlation ≥ 0.92 (5 runs) PA (2018) ns ±10% ±5% (3 runs) PR (2019) 1 in./mi ±5% ±5% (3 runs) RI (2018) Vendor certified Vendor certified Vendor certified SC (2018) 1 in./mi ±5% ±5% (5 runs) SD (2018) 1 in./mi ±5% ±5% (5 runs) TN (2018) Vendor certified Vendor certified Vendor certified TX (2018) 1 in./mi ≤ DS im/.ni 6 ≤ 3 in./mi UT (2018) ns SD < 10% historical to current <5% (10 runs) VT (2018) 1 in./mi ±5% ±5% (3 runs) WA (2018) ns Cross-correlation ≥ 95% Cross-correlation ≥ 92% WI (2018) na >90% of known profile >92% multiple runs WV (2019) 0.1 in./mi ±5% ±5% aThe year of publication is provided as a reference. Note: ns = not specified; avg. = average, na = not applicable, and SD = standard deviation. Table 15. Summary of agency criteria for equipment calibration and certification: IRI.

Literature Review 27   Statea Resolution Accuracy (to reference value) Repeatability AK (2018) 0.02 in. ≥ snur 01 fo %5 < DS %59 AL (2018) ns ±0.1 in., 95% of segments Avg. ± 0.05 in. (5 runs) AR (2018) 0.01 in. ±0.05 in. SD ± 0.05 in. (3 runs) AZ (2018) 0.0001 in. Contractor provided Contractor provided CA (2018) 0.05 in. ±0.1 in. SD ± 0.06 in. (3 runs) CT (2018) ≤ )snur 5( .ni 60.0 ± DS %80.0± .ni 40.0 DC (2019) 0.01 in. ±0.04 in. ±0.08 in. DE (2018) 0.01 in. avg. or max ±0.05 in. SD < 5 ft; SD < 0.05 in. (10 runs) FL (2018) 0.01 in. ±0.06 in. ±0.06 in. (3 runs) GA (2018) 0.01 in. ±0.5 mm vendor baseline value ±0.06 in. (3 runs) IA (2018) 0.01 in. ±0.05 in. ±0.06 in. (3 runs) ID (2018) 0.01 in. ±0.06 in. ±0.06 in. (3 runs) IL (2018) Vendor certified Vendor certified Vendor certified IN (2019) 0.01 in. ±0.05 in. ±0.06 in. (3 runs) MA (2018) 0.01 in. >85% SD < 0.04 in. (10 runs) & historical MD (2018) ns Based on 40 sections Based on 40 sections ME (2018) ≤ )LWP 59( )snur 5( .ni 1.0± naeM .ni 21.0± .ni 40.0 MO (2018) ±0.01 mm ±0.06 in. ±0.06 in. (3 runs) MS (2019) 0.01 in. ±0.06 in. ±0.006 in. MT (2018) 0.01 in. ±0.06 in. ±0.06 in. (3 runs) NC (2018) 0.01 in. Bias: <5% Precision: <5% ND (2018) 0.01 in. ±0.019 in. reference ±0.06 in. (3 runs) NE (2018) 0.254 mm ± 0.4826 mm ±1.524 mm (3 runs) NH (2019) 0.02 in. ±0.12 in. (5 runs) ±0.12 in. max. variance threshold NJ (2019) 0.01 in. ±0.06 in. ±0.06 in. (3 runs) NM (2018) 0.4 in. >85% SD < 0.04 in. (10 runs) & historical NV (2018) 0.01 in. na ns NY (2018) 0.01 in. ns ns OK (2018) 0.01 in. ±0.08 in. ±0.08 in. (3 runs) OR (2018) 0.01 in. ±0.10 in. ±0.05 in. (3 runs) PA (2018) ns ±10% ±5% (3 runs) PR (2019) 0.01 in. ±0.5 mm ±0.06 in. (3 runs) RI (2018) Vendor certified Vendor certified Vendor certified SC (2018) 0.01 in. ±0.08 in. ±0.08 in. (5 runs) SD (2018) 0.01 in. ±0.06 in. ±0.06 in. (5 runs) TN (2018) Vendor certified Vendor certified Vendor certified TX (2018) 0.05 in. ±0.1 in. SD ± 0.06 in. UT (2018) ns SD < 2 mm previous verification < 1 mm (10 runs) VT (2018) 0.01 in. ±0.06 in. ±0.06 in. (3 runs) WA (2018) ns ±0.08 in. of manual survey CV < 10% (3 runs) WI (2018) na >90% of known profile >92% multiple runs WV (2019) 0.01 in. ±0.06 in. ±0.06 in. aThe year of publication is provided as a reference. Note: avg. = average, CV = coefficient of variation, max. = maximum, na = not applicable, ns = not specified, PWL = percent within limits, and SD = standard deviation. Table 16. Summary of agency criteria for equipment calibration and certification: rut depth.

28 Automated Data Collection and Quality Management for Pavement Condition Reporting Statea Resolution Accuracy (to reference value) Repeatability AK (2018) na na na AL (2018) ns ns ±0.01 in. AR (2018) 0.01 in. ±0.05 in. SD ± 0.05 in. (3 runs) AZ (2018) 0.0001 in. Contractor provided Contractor provided CA (2018) 0.05 in. ±0.1 in. SD ± 0.06 in. (3 runs) CT (2018) ≤ )snur 5( .ni 60.0 ± DS %80.0± .ni 40.0 DC (2019) 0.01 in. ±0.06 in. ±0.06 in. reference value DE (2018) 0.01 in. ±50 count (10 runs) SD < 5 count (10 runs) FL (2018) 0.01 in. ±0.08 in. ±0.08 in. (3 runs) GA (2018) 0.01 in. ±0.06 in. vendor baseline value ±0.06 in. for 3 runs IA (2018) 0.01 in. ±0.05 in. ±0.06 in. (3 runs) ID (2018) 0.01 in. ±0.06 in. ±0.06 in. (3 runs) IL (2018) Vendor certified Vendor certified Vendor certified IN (2019) 0.01 in. ±0.06 in. ±0.06 in. (3 runs) MA (2018) 0.04 in. ±0.05 in. per severity SD < 15% (10 runs) & historical MD (2018) ns Based on 40 sections Based on 40 sections MI (2018) Vendor certified ns ns MO (2018) ±0.01 mm ±0.004 in. ±0.04 in. (3 runs) MS (2019) 0.01 in. ±0.06 in. ±0.08 in. MT (2018) 0.01 in. ±0.06 in. ±0.06 in. (3 runs) NC (2018) 0.01 in. ±0.06 in. <5% ND (2018) 0.01 in. ±0.06 in. ±0.06 in. (3 runs) NE (2018) 0.254 mm ±1.524 mm (3 runs) NJ (2019) 0.01 in. ±0.06 in. ±0.06 in. (3 runs) NM (2018) 0.04 in. ±0.05 in. SD < 15% (10 runs) & historical NV (2018) 0.01 in. na ns NY (2018) 0.01 in. ns ns OK (2018) ±0.01 in. ±0.04 )snur 3( .ni 40.0± .ni OR (2018) 0.01 in. ±0.06 in. ±0.06 in. (3 runs) PA (2018) ns ±10% ±5% (3 runs) PR (2019) 0.01 in. ±0.06 in. ±0.06 in. (3 runs) RI (2018) Vendor certified Vendor certified Vendor certified SC (2018) 0.01 in. ±0.08 in. ±0.08 in. (5 runs) SD (2018) 0.01 in. ±0.06 in. ±0.06 in. (5 runs) TN (2018) Vendor certified Vendor certified Vendor certified TX (2018) 0.001 in. na na WA (2018) ns ±0.08 in. CV < 10% (3 runs) WV (2019) 0.01 in. ±0.06 in. ±0.06 in. aThe year of publication is provided as a reference. Note: CV = coefficient of variation, na = not applicable, ns = not specified, and SD = standard deviation. ±1.524 mm Table 17. Summary of agency criteria for equipment calibration and certification: faulting.

Literature Review 29   Statea Resolution Accuracy (to reference value) Repeatability AK (2018) na ±5% SD < 5% (10 runs) AL (2018) ns ns 0.1 ft/0.01 mile AR (2018) 1% ±20% or ±3 points ±20% or ±3 points (3 runs) AZ (2018) 1% Contractor provided Contractor provided CA (2018) na ±10% na CT (2018) 1% HPMS & 0.1 ft agency ±30% asphalt, ±20% concrete CV < 30% asphalt, ±20% concrete (5 runs) DC (2019) na ±10% na DE (2018) 1 ft2 per severity ±50 ft2 (10 runs) SD < 5 ft2 (10 runs) FL (2018) 1% ±20% ±20% (3 runs) GA (2018) Varies ±20% vendor baseline value na IA (2018) Varies ±10% na ID (2018) na ±10% na IL (2018) Vendor certified Vendor certified Vendor certified IN (2019) Varies ±20% na MA (2018) na ±10% asphalt, ±5% concrete SD < 15% (10 runs) & historical MD (2018) ns Based on 40 sections Based on 40 sections ME (2018) 1 mm ±3 mm SD ± 3 mm (5 runs) MI (2018) Vendor certified ns ns MO (2018) 2 mm ±10% na MS (2019) 1% ±20% ±20% MT (2018) Varies ±20% na NC (2018) Varies ±20% ns ND (2018) Varies ±20% na NE (2018) Varies ±20% na NH (2019) 33% cracks ≤ 1/8-in. wide, 60% cracks 1/8- to 1/4-in. wide, 85% cracks ≥ 1/4-in. wide sn sn NJ (2019) 0.1 mi 95% na NM (2018) Varies ±5% asphalt, ±5% concrete SD < 15% (10 runs) & historical NV (2018) Varies na ns NY (2018) 1% ns ns OH (2018) ns ns ns OK (2018) 1% ±10% asphalt, ±20% concrete ±10% (3 runs) OR (2018) na ±20% na PA (2018) ns ±10% ±5% (3 runs) PR (2019) Varies ±20% na RI (2018) Vendor certified Vendor certified Vendor certified SC (2018) Varies ±10% na SD (2018) ns ns ns TN (2018) Vendor certified Vendor certified Vendor certified TX (2018) na ≤15 points ≥90% compliance UT (2018) ns SD < 15% previous validation <15% (10 runs) VT (2018) ns ns ns WA (2018) ns >90% manual to automated ns WI (2018) ns ns ns WV (2019) 0.10% ±10% ±10% aThe year of publication is provided as a reference. Note: Cracking may include percent cracking or individual distress types to determine percent cracking. CV = coefficient of variation, na = not applicable, ns = not specified, and SD = standard deviation. Table 18. Summary of agency criteria for equipment calibration and certification: percent cracking.

30 Automated Data Collection and Quality Management for Pavement Condition Reporting Statea IRI Rutting AK (2018) SD < 5% 10 runs SD < 0.04 in. 10 runs AL (2018) Avg. of 5 runs for target during production; <5% single lane; single pass weekly over 1 control site; cross-correlation ≥ 88% ±0.1 in. reference value AR (2018) ±10% historical values 95% PWL ±0.05 in. historical values 95% PWL AZ (2018) Confidence interval (10 runs, 5 sites) Weekly verification (1 run, 1 site) Confidence interval (10 runs, 4 sites) Weekly verification (3 run, 4 sites) CA (2018) SD ≤ 5% (3 runs) ±10% agency value SD ≤ 0.06 in. (3 runs) ±0.06 in. agency value CT (2018) SD ≤ 5% (5 runs) Weekly left & right IRI values differ ≤ 50 in./mi SD ≤ 0.40 in. (5 runs) Weekly left & right rutting values differ ≤0.25 in. DC (2019) ≥20 in./mi, ≤400 in./mi on Interstate ≥30 in./mi non-Interstate ≤0.75 in. DE (2018) Left & right IRI ≤ 50 in./mi Left & right values ≤ 0.25 in. FL (2018) Repeatability score ≥92 Accuracy score ≥90 ns GA (2018) ±5% previous year ±5% previous year IA (2018) ±10% reference value SD 5% (3 runs) ±0.1 in. reference value SD 5% (3 runs) ID (2018) SD ≤ 5% (10 runs) SD ≤ 10% historical average SD ≤ 0.4 in. (10 runs) SD ≤ 0.4 in. historical average IL (2018) ±10% of initial run SD ≤ 5% (10 runs) ±0.08 in. of initial run SD ≤ 0.04 in. (10 runs) IN (2019) SD ≤ 5% (10 runs) SD ≤ 10% historical average SD ≤ 0.4 in. (10 runs) SD ≤ 0.4 in. historical average MA (2018) SD ≤ DS )snur 01( %5 ≤ 0.4 in. (10 runs) MD (2018) ±previous year results ±previous year results ME (2018) AASHTO Standard AASHTO Standard MI (2018) 5 run avg. ± reference value 5 run avg. ± reference value MN (2018) ±5% reference value Max. ± 0.10 in. reference value MO (2018) 95% compliance with standards 95% compliance with standards MS (2019) CV < 5% (5 runs) ±0.04 in. (5 runs) MT (2018) 95% compliant with standards 95% compliant with standards NC (2018) Data consistency, year-to-year change Data consistency, year-to-year change ND (2018) 95% compliance with standards 95% compliance with standards NE (2018) 95% compliance with standards 95% compliance with standards NH (2019) Repeat runs compared with reference value Repeat runs compared with reference value NJ (2019) Data checks Data checks NM (2018) SD ≤ 10% historical average, SD ≤ 5% (10 runs) SD ≤ 0.4 in. (10 runs) & historical average NV (2018) 95% compliance with standards 95% compliance with standards NY (2018) ±10% reference value ±0.13 in. reference value OH (2018) Data checks Data checks OK (2018) Completeness, accuracy, repeatability Completeness, accuracy, repeatability OR (2018) Cross-correlation ≥ 92% (5 runs) ±0.05 in. (3 runs) PA (2018) Control, blind site, & 2.5% random sample Control, blind site, & 2.5% random sample PR (2019) 95% compliance 95% compliance with standards RI (2018) ±10% (3 runs) Avg. ± 0.13 in. SC (2018) Verification testing Verification testing SD (2018) Flagged if running average > ±10% Flagged if running average ± 0.06 in. TN (2018) t-test reference vs. automated (5 runs) t-test reference vs. automated (5 runs) TX (2018) >3 in./mi difference left & right wheel path >0.06 in. difference left & right wheel path UT (2018) SD ≤ 10% of last validation, SD ≤ 5% (10 runs) SD ≤ 2 mm of last validation SD ≤ 1 mm (10 runs) VT (2018) ±5% (3 to 5 runs) Absolute value 1.000 in./mi >5% with no rehabilitation & >15% historical ±5% (3 to 5 runs) Absolute value 0.050 in. >5% with no rehabilitation & >15% historical WA (2018) ±20 in./mi 4-week moving average SD < 8.5 in./mi last 5 weeks ±0.08 in. 4-week moving average SD < 0.025 in. last 5 weeks WI (2018) Check any deviations ±10% Check any deviations ±0.1 inch WV (2019) 95% verification testing 95% verification testing aThe year of publication is provided as a reference. Note: avg. = average, CV = coefficient of variation, IRI = international roughness index, max. = maximum, PWL = percent within limits, and SD = standard deviation. Table 19. Summary of agency criteria for quality control: IRI and rut depth.

Literature Review 31   Statea Faulting Cracking AK (2018) na SD < 15% 10 runs AL (2018) ±0.1 reference value ≥ 95% samples 3% sample reference value, Pearson’s r correlation AR (2018) 95% PWL ± 0.05 in. historical values 95% PWL ± 20% historical values AZ (2018) Confidence interval (10 runs, 1 site) Weekly verification (3 runs, 1 site) Acceptable range 1 run & 3 manual ratings (6 sites) & monthly 1 run & 3 manual ratings (6 sites) CA (2018) SD ≤ 0.06 in. (3 runs), ±0.06 in. agency SD ≤ 15% multiple runs or historical value CT (2018) Weekly >0 & ≤ DS .ni 0.1 ≤ 20% length (5 runs) DC (2019) ≤ sn .ni 57.0 DE (2018) ≤1 in. ≤100% FL (2018) ns ns GA (2018) ns ns IA (2018) ±0.1 in. reference value; SD 5% (3 runs) ±10% reference value ID (2018) Values >0 and ≤1 in. DS ≤ 15% (10 runs) & SD ≤ 15% historical avg. IL (2018) ns ±15% of initial run & SD ≤ 15% (10 runs) IN (2019) Value >0 and ≤ DS .ni 1 ≤ 15% (10 runs) & SD ≤ 15% historical avg. MA (2018) SD < 15% (10 runs) and ±0.05 in. per severity SD ≤15% (10 runs) & ±10% area (asphalt) & ±5% total area (concrete) MD (2018) ±previous year results Image quality ME (2018) na 100% visual review MI (2018) 5-run avg. ± reference value 5 run avg. ± reference value MN (2018) Avg. ± 0.05 in. reference value Image must show 90% of cracks MO (2018) 95% compliance with standards 95% compliance with standards MS (2019) ±0.04 in. (5 runs) CV < 15% (5 runs) MT (2018) 95% compliant with standards 80% match, manual to automated NC (2018) Data consistency, year-to-year change Data consistency, year-to-year change ND (2018) 95% compliance with standards 80% match, manual to automated NE (2018) 95% compliance with standards To be determined NH (2019) na ns NJ (2019) Data checks Data checks NM (2018) ≤ DS .ni 1 ≤ 15% (10 runs) & historical avg. NV (2018) 95% compliance with standards 95% compliance with standards NY (2018) ±0.05 in. reference value ±10% reference value OH (2018) Data checks Data checks OK (2018) Completeness, accuracy, repeatability Completeness, accuracy, repeatability OR (2018) ±0.06 in. (3 runs) SD ≤ 10% (3 runs and/or historical avg.) PA (2018) Control, blind, & 2.5% random sample Control, blind site, & 2.5% random sample PR (2019) 95% compliance with standards 80% match, manual to automated RI (2018) avg. ± 0.05 in. ±10% 90% of the time SC (2018) Verification testing Verification testing SD (2018) Flagged if running average ±0.06 in. Flagged if tolerance >1% TN (2018) t-test reference vs. automated (5 runs) t-test reference vs. automated (5 runs) TX (2018) ns ns UT (2018) ns SD ≤ 15% of last validation, SD ≤ 15% (10 runs) VT (2018) ±5% (3 to 5 runs) ns WA (2018) ns 90% of t-test & R2 WI (2018) ns Check any deviations ±10% WV (2019) 95% verification testing 95% verification testing aThe year of publication is provided as a reference. Note: avg. = average, CV = coefficient of variation, na = not applicable, ns = not specified, PWL = percent within limits, and SD = standard deviation. Table 20. Summary of criteria for agency quality control: faulting and cracking.

32 Automated Data Collection and Quality Management for Pavement Condition Reporting Statea IRI Rutting AK (2018) 95% verification testing 95% verification testing AL (2018) Based on control site Based on control site AR (2018) 95% ± 10% historical value (5% sample) 95% ± 0.05 in. historical value (5% sample) AZ (2018) 95% weekly control site & 5%–10% sample 95% weekly control site & 5%–10% sample CA (2018) 95% ± 10% reference value 95% ± 0.06 in. reference value CT (2018) 95% monthly verification sites 95% monthly verification sites DC (2019) 95% verification testing 95% verification testing DE (2018) ≥90% within normal bounds ≥90% within normal bounds FL (2018) 90% certification & verification site tests ns GA (2018) 95% verification testing 95% verification testing HI (2019) 95% ± 5% of ground truth 95% ± 0.1 in. of ground truth IA (2018) 95% verification testing 95% verification testing ID (2018) 95% verification testing 95% verification testing IL (2018) ±10% historical data & random sample ±0.08 in. historical data & random sample IN (2019) 95% verification testing 95% verification testing MA (2018) 95% verification testing 95% verification testing MD (2018) > ±2% change from previous year > ±2% change from previous year ME (2018) 20 to 900 in./mi 0–1.5 in. MI (2018) Percent change from historical data (2% sample) Percent change from historical data (2% sample) MN (2018) Daily & periodic skcehc cidoirep & yliaD skcehc MO (2018) 95% weekly comparisons to previous years MS (2019) 90% previous year 90% compliant 5% sample inspection MT (2018) 95% verification testing 95% verification testing NC (2018) ns ns ND (2018) 95% compliance with standards 95% compliance with standards NE (2018) ±5% baseline, monthly, daily submittal ±5% baseline, monthly, daily submittal NH (2019) Data checks Data checks NJ (2019) 95% PWL 5% sample 95% PWL 5% sample NM (2018) 95% verification testing 95% verification testing NV (2018) 95% verification testing 95% verification testing NY (2018) Random sample & historical trend Random sample & historical trend OH (2018) 100% data check, historical reference All high values are reviewed against imagery OK (2018) 95% data checks 95% data checks OR (2018) 100% verification, 95% sample 100% verification, 95% sample PA (2018) 95 PWL ± 25% 90% PWL ± 20% PR (2019) 95% verification testing 95% verification testing RI (2018) Random sample, F- and t-test Random sample, F- and t-test SC (2018) 95% compliance weekly inspection 95% compliance weekly inspection SD (2018) 100% data check, historical reference 100% data check, historical reference TN (2018) Verify data meets standards Verify data meets standards TX (2018) 100% verification testing 100% verification testing UT (2018) Consistency & compare with historical Consistency & compare with historical VT (2018) 5% random sample low & high quartile 5% random sample low & high quartile WA (2018) Verification te gnitset noitacifireV gnits WI (2018) Compare to histori lacirotsih ot erapmoC lac WV (2019) 95% weekly control sites, ≤10% sample 95% weekly control sites, ≤10% sample aThe year of publication is provided as a reference. Note: ns = not specified; PWL = percent within limits. 95% weekly comparisons to previous years Table 21. Summary of agency acceptance: IRI and rutting.

Literature Review 33   Statea Faulting Cracking AK (2018) na 95% verification testing AL (2018) Based on control site Based on control site AR (2018) 95% ± 0.05 in. historical (5% sample) 95% ± 20% historical (5% sample) AZ (2018) 95% weekly control site & 5%–10% sample 90% weekly control site & 5%–10% sample CA (2018) 95% ± 0.06 in. reference value 85% ± 10% reference value CT (2018) 95% monthly verification sites 95% monthly verification sites DC (2019) 95% verification testing 95% verification testing DE (2018) ≥90% within normal bounds ≥90% within normal bounds FL (2018) 80% certification & verification site tests 80% certification & verification site tests GA (2018) 95% verification testing 80% match, manual vs. automated HI (2018) na 90% ± 15% ground truth IA (2018) 95% verification testing >90% of 5% sample, accurate ID (2018) 95% verification testing 90% verification testing IL (2018) Reasonable to historical & random sample Reasonable to random sample IN (2019) 95% verification testing 80% match, manual vs. automated MA (2018) 95% verification testing 95% verification testing MD (2018) > ±2% change from previous year >1% previous year (asphalt); 5% manual (concrete) ME (2018) na 0%–60% MI (2018) Within manual fault measurement Agency sample manual review, compare with previous MN (2018) Daily & periodic checks Manual review 10% of segments MO (2018) 95% weekly comparisons with previous years 95% weekly comparisons with previous years MS (2019) 90% compliant 5% sample inspection 90% compliant 5% sample inspection MT (2018) 95% verification testing 80% verification testing NC (2018) ns 90% PWL random sections ND (2018) 95% compliance with standards 80% match, manual vs. automated NE (2018) ±5% baseline; monthly, daily inspection ±10 points, manual to automated rating NH (2019) Data checks Data checks NJ (2019) 95% PWL 5% sample 95% PWL 5% sample NM (2018) 95% verification testing 95% verification testing NV (2018) 95% verification testing 95% verification testing NY (2018) Random sample & historical trend Random sample & historical trend OH (2018) All high & zero values image review All high & zero values image review OK (2018) 95% data checks 95% data checks OR (2018) 100% verification, 95% expected values 100% verification, compare with previous < ±10 pts PA (2018) ns 90 PWL ± 20% PR (2019) 95% verification testing 80% match, manual vs. automated RI (2018) Random sample, F- & t-test Random sample, F- & t-test SC (2018) 95% compliance weekly inspection 95% compliance 25% weekly inspection SD (2018) 100% data check, historical reference 100% data check, historical reference TN (2018) Verify data meets standards Verify data meets standards TX (2018) ns 90% compliance 6% sample UT (2018) Consistency & compared to historical Consistency & compared to historical VT (2018) ns 5% random sample lowest & highest quartile WA (2018) Verification testing 5% sample, ≥90% (R2 & t-test) WI (2018) Historical ≤ −0.05 in. & ≥0.1 in. Historical ≤ −5% & ≥10% WV (2019) 95% weekly control sites, ≤10% sample 95% weekly control sites, ≤10% sample aThe year of publication is provided as a reference. Note: na = not applicable, ns = not specified, PWL = percent within limits. Table 22. Summary of agency acceptance: faulting and cracking.

34 Automated Data Collection and Quality Management for Pavement Condition Reporting Summary of Chapter 2 Over the past several decades, SHAs have been transitioning from manual to semiautomated to fully automated pavement condition surveys. This transition was originally initiated through collection of transverse and longitudinal profiles for determining IRI, rut depth, and faulting. Automated quantification of surface distress was initiated by capturing video of the pavement surface, which transitioned to digital images, then to 2D technology, and finally, to 2D and 3D technology. However, a number of challenges in transitioning to an APCS have been noted: • Lack of method standardization, • Lack of information related to how agencies successfully transitioned to an APCS, • Potential issues on compatibility with historical records, • Impact on historical performance models, • Cost and time to conduct the APCS, and • Data quality concerns. The primary benefits of transitioning from manual surveys to an APCS include improved effi- ciency and safety. Today, the majority of SHAs conduct APCS by using 2D and 3D technology. SHAs are required to collect, analyze, and submit pavement condition measures to several national programs, including PM2, HPMS, and GASB-34. For PM2 reporting, SHAs are required to submit IRI, rutting, faulting, and percent cracking annually for the Interstate highway system and every 2 years for the non-Interstate NHS. For low-speed highways (<40 mph), agencies may submit PSR values in lieu of IRI, faulting, cracking, and percent cracking. For HPMS reporting, SHAs are required to submit the same pavement condition measures along with other report- ing requirements. Finally, for GASB-34, SHAs report general infrastructure assets along with depreciation or preservation cost. As required by PM2, SHAs are required to develop, submit, and use a DQMP for the pave- ment condition survey, regardless of the type of survey method. In general, the DQMP includes criteria for equipment calibration and certification, quality control, and acceptance criteria. Of the 48 SHA DQMPs reviewed, all include criteria for IRI, rut depth, faulting, and percent crack- ing, as required by PM2. In addition, agencies include criteria for other distress types (e.g., block cracking, patching, raveling), profile measures (e.g., cross slope, macrotexture), and image (e.g., clarity, perspective). This chapter also summarizes the criteria for equipment calibration and certification, data quality control, and acceptance for IRI, rut depth, faulting, and cracking. Component Criteria IRI Rutting Faulting Cracking Equipment calibration and certification Resolution 1 in./mi 0.01 in. 0.01 in. Varies Accuracy to reference value ±5% ±0.06 in. ±0.06 in. ±20% Repeatability (multiple runs) ±5% ±0.06 in. ±0.06 in. na Data quality control DS an ≤ 5% ±0.04 in. 95% compliant with standards SD < 15% of multiple runs and/or historical values Data acceptance Compliance with verification testing 95% 95% 95% 95% Note: na = not applicable Table 23. Summary of common requirements for APCS data quality.