A GUIDE TO COLLECTING, PROCESSING, AND MANAGING ROADWAY ASSET INVENTORY DATA FINAL REPORT

NCHRP Project 20 07/Task 357 A GUIDE TO COLLECTING, PROCESSING, AND MANAGING ROADWAY ASSET INVENTORY DATA FINAL REPORT Requested by: American Association of State Highway and Transportation Officials (AASHTO) Standing Committee on Highways Subcommittee on Maintenance Prepared by: Kathryn A. Zimmerman, P.E. Kartik Manda Applied Pavement Technology, Inc. 115 West Main Street, Suite 400 Urbana, Illinois 61801 June 2015 The information contained in this report was prepared as part of NCHRP Project 20-07, Task 357, National Cooperative Highway Research Program. SPECIAL NOTE: This report IS NOT an official publication of the National Cooperative Highway Research Program, Transportation Research Board, National Research Council, or The National Academies.

ACKNOWLEDGMENTS This study was requested by the American Association of State Highway and Transportation Officials (AASHTO), and conducted as part of the National Cooperative Highway Research Program (NCHRP) Project 20-07. The NCHRP is supported by annual voluntary contributions from the state Departments of Transportation. Project 20-07 provides funding for quick response studies on behalf of the AASHTO Standing Committee on Highways. The report was prepared by Applied Pavement Technology, Inc. The work was guided by a task group which included Tanveer Chowdhury, Virginia DOT; William D. Bill Drake, Jr, Louisiana DOTD; Christopher C. Harris, Tennessee DOT; Thomas J. Kazmierowski, Golder Associates; Mary A. Martini, Nevada DOT; Roger E. Smith, Texas A&M University (retired); Lonnie R. Watkins, North Carolina DOT; and Nastaran Saadatmand, FHWA. The project manager was Amir N. Hanna, NCHRP Senior Program Officer. DISCLAIMER The opinions and conclusions expressed or implied are those of the research agency that performed the research and are not necessarily those of the Transportation Research Board or its sponsoring agencies. This report has not been reviewed or accepted by the Transportation Research Board Executive Committee or the Governing Board of the National Research Council.

NCHRP Project 20-07/Task 357 A GUIDE TO COLLECTING, PROCESSING, AND MANAGING ROADWAY ASSET INVENTORY DATA FINAL REPORT Requested by: American Association of State Highway and Transportation Officials (AASHTO) Standing Committee on Highways Subcommittee on Maintenance Prepared by: Kathryn A. Zimmerman, P.E. Kartik Manda Applied Pavement Technology, Inc. 115 West Main Street, Suite 400 Urbana, Illinois 61801 June 2015 The information contained in this report was prepared as part of NCHRP Project 20-07, Task 357, National Cooperative Highway Research Program. SPECIAL NOTE: This report IS NOT an official publication of the National Cooperative Highway Research Program, Transportation Research Board, National Research Council, or The National Academies.

Final Report TABLE OF CONTENTS CHAPTER 1 INTRODUCTION... 1 PROJECT OVERVIEW... 1 RESEARCH SCOPE AND TASKS... 1 Task 1: Summarize the State of the Practice... 1 Task 2: Identify Trends... 2 Task 3: Develop Guidance... 2 Task 4: Prepare Documentation... 2 DISTRIBUTION OF RESEARCH PRODUCTS... 3 CHAPTER 2 SUMMARY OF PRACTICE... 4 INTRODUCTION... 4 TECHNOLOGY BEING USED TO ESTABLISH ASSET INVENTORIES... 4 AVAILABLE REFERENCES ON BUILDING AN ASSET INVENTORY... 5 Data Collection Techniques... 5 LiDAR... 8 Data Quality... 9 STATUS OF ASSET INVENTORIES IN STATE DOTS... 10 Drainage Assets... 11 Roadside Assets... 11 Pavement and Bridge Assets... 12 Traffic Assets... 12 Special Facilities... 13 EMERGING TRENDS... 14 EMERGING TECHNOLOGY... 15 360-Degree Camera... 15 Flash LiDAR... 15 Airborne LiDAR... 15 Driverless Cars... 16 CHAPTER 3 CONCLUSIONS... 17 FUTURE RESEARCH NEEDS... 17 REFERENCES... 19 i

Final Report LIST OF FIGURES Figure 1. Inventory status of drainage assets (NCHRP 2015)... 11 Figure 2. Inventory status of roadside assets (NCHRP 2015)... 12 Figure 3. Inventory status of pavement assets (NCHRP 2015)... 12 Figure 4. Inventory status of traffic assets (NCHRP 2015)... 13 Figure 5. Inventory status of special facilities (NCHRP 2015)... 13 LIST OF TABLES Table 1. Suitability of different methods of data collection (Zimmerman and Stivers 2007)... 6 ii

Final Report AUTHOR ACKNOWLEDGMENTS The research report herein was performed under NCHRP Project 20-07, Task 357 by Applied Pavement Technology, Inc. (APTech). Ms. Kathryn A. Zimmerman, P.E., served as the Principal Investigator for this study. She was assisted by Mr. Kartik Manda, an Engineering Associate at APTech. ABSTRACT This project was initiated by the National Cooperative Highway Research Program to develop guidance for establishing and managing roadway asset inventories. The resulting Guide, which was written as a standalone document, can be used by transportation agencies to help make informed decisions on the type of technology most appropriate for collecting asset inventory information and the considerations that must be taken into account for processing and managing the data. The study concentrated on both manual and automated data collection approaches, including manual surveys, photogrammetric methods, and remote sensing technology (e.g., mobile LiDAR). The Guide includes considerations that should be evaluated during all phases of establishing or updating an asset inventory. First, the Guide addresses technical considerations that should be taken into account regardless of the data collection selected, such as developing criteria for classifying assets and developing data collection standards. Secondly, the Guide presents factors to consider in determining the appropriateness of each of the three technologies used in collecting inventory data. This section includes factors such as the level of accuracy required and the visibility of the asset from the road. Next, the Guide includes considerations for collecting the data, including differences depending on whether the data will be collected using in-house personnel or an outside contractor. Finally, the Guide suggests considerations for managing the data effectively, including topics such as storage requirements and update schedules. iii

Final Report PROJECT OVERVIEW CHAPTER 1 INTRODUCTION Over the past decade and since the passage of recent legislation (commonly known as Moving Ahead for Progress in the 21 st Century Act or MAP-21), there has been an increasing emphasis on the use of performance data to drive agency investment decisions as part of a comprehensive asset management program. While inventory and performance data has been collected on pavement and bridge assets for many years, there is less consistency in the status of roadway asset inventories for other assets such as guardrails, culverts, and signs. National Cooperative Highway Research Program (NCHRP) Project 20-07, Task 357 was initiated in 2014 to develop guidance for establishing and managing roadway asset inventories among state Departments of Transportation using current technology. This document represents the Final Report for the project. It summarizes the project activities and documents the assessment of current practice that was conducted during the early stages of the project. The information gathered from this activity served as the basis for developing the Guide, which is presented as a standalone document as an attachment to this Final Report. The Guide considers both manual and automated technologies and includes factors that highway agencies should consider when deciding which approach to use for building its asset inventory. Once the decision is made, the Guide includes recommendations for making the best use of the technology for collecting, processing, and managing roadway asset inventory data, such as guardrails, tower lighting, signs, and drainage features. It is important to note that the Guide does not address the performance criteria that are often used to monitor the level of service being provided to the traveling public or to prepare maintenance budgets. Although the Guide focuses primarily on building, maintaining, and managing asset inventories, the same technology can often be used to evaluate asset performance. As a result, many of the same considerations identified for establishing an asset inventory are relevant to the process of assessing the condition of these assets. RESEARCH SCOPE AND TASKS The project objective was to develop practical guidance that could be used by highway agency practitioners for collecting, processing, and managing roadway asset inventory data. This objective was accomplished through the completion of the four tasks described below. Task 1: Summarize the State of the Practice The project began with a series of activities designed to provide a good understanding of the state of the practice. One of the activities included a literature search of the readily available documentation on collecting, processing, and managing roadway asset inventories. One of the major sources of information included the results from NCHRP Synthesis 470, titled Maintenance Quality Assurance Field Inspection Practices (NCHRP 2015). The preparation of the synthesis included the conduct of a survey into the practices in state highway agencies for collecting inventory information on a variety of different types of assets (e.g., culverts, sidewalks, fences, pavement shoulders, and signs). This information proved to be very useful in determining the status of asset inventories and the methods used to collect the information. Information from the synthesis is included in Chapter 2 of this report. In addition to reviewing reports and other forms of documentation, the project team conducted interviews with both data collection vendors working in the state data collection market and state 1

Final Report DOT practitioners who have used different forms of technology to establish asset inventories. The information from the interviews influenced the development of the Guide and some of the information is used to illustrate the considerations identified. At the conclusion of Task 1, a summary report was prepared and distributed to the project s Technical Panel for review and comment. The feedback provided by the Technical Panel was also instrumental in shaping the Guide s content. Task 2: Identify Trends The information obtained during Task 1 served as the basis for identifying trends in the asset inventory information being collected by state agencies and the methodologies being used. Additionally, the interviews with state practitioners provided a good understanding of the factors that influenced the selection of a technology for building their asset inventories. The states selected to be interviewed represented a range of data collection methodologies, including both manual and automated approaches. The trends that were observed are incorporated into the Guide. Task 3: Develop Guidance During Task 3, the research team used the information obtained during Tasks 1 and 2 to develop the framework for creating the Guide. The Guide, which is presented as an attachment to this Final Report, addresses the following four steps associated with collecting, processing, and managing asset inventory data: Step 1: Getting ready to select a methodology This step includes the organizational issues that need to be addressed to assess an agency s needs. This step involves deciding what assets to include in the inventory, identifying the users of the data, determining the level of detail needed, and establishing the characteristics that will be used to describe each asset. Step 2: Selecting a methodology Using the information obtained during step 1, the second step involves selecting the most appropriate methodology to meet the agency s needs. The decision is based on a number of different factors, related to the visibility of the asset from the road, the level of detail needed, safety considerations, the potential for collaboration with other data collection activities, and available resources. Step 3: Collecting the data Immediately before and during the data collection processes, steps need to be taken to ensure the quality of the data. This step includes the activities involved in securing a data collection vendor (if appropriate), establishing test sites to verify the technology provides the necessary data, and monitoring the quality of the data throughout the data collection process. Step 4: Processing and managing the data The final step involves processing the data to extract the necessary information and ensuring that the data is updated on a regular cycle. This step contains the factors that must be taken into consideration to ensure the best possible use of the information within the agency. Task 4: Prepare Documentation The last project task involved preparing this Final Report, which includes the guidance described earlier. 2

Final Report DISTRIBUTION OF RESEARCH PRODUCTS This document was developed primarily for maintenance personnel in state DOTs responsible for developing and maintaining a roadway asset inventory. It is designed to assist these individuals in determining the type of technology most appropriate for building and maintaining the inventory, the technical and organizational considerations that should be addressed prior to building the inventory, and the data processing and management issues that should be addressed with each of the different forms of technology. The considerations described in the Guide are not unique to practices in state DOTs; therefore, the information provided in this document can be equally useful to maintenance personnel in cities, counties, or other transportation agencies. In addition to maintenance personnel, other practitioners may benefit from the information provided in this Guide. For instance, the information may help an agency that is using automated equipment for pavement management data collection find new uses for the digital images that are being collected. Similarly, an agency that is using a vehicle equipped with LiDAR for collecting inventory information may discover new applications for the technology to support the agency s design activities. In addition to making this report available through the NCHRP website, the information contained in this document will be distributed to practitioners through technical presentations at meetings such as the Transportation Research Board (TRB) Annual Meeting and meetings of the American Association of State Highway and Transportation Officials (AASHTO) Subcommittee on Maintenance. Opportunities to present the information through webinars and/or workshops will also be sought. 3

Final Report INTRODUCTION CHAPTER 2 SUMMARY OF PRACTICE An agency s ability to make sound, defensible investment decisions relies in part on the availability of a comprehensive asset inventory, a method of assessing current conditions and performance, and tools for evaluating the impacts of different investment strategies on network performance. Establishing an inventory is a fundamental step in establishing an asset management program. This chapter introduces the manual and automated technologies that are commonly used to establish asset inventories and documents the use of the technology in practice. The chapter also summarizes the status of asset inventories in state highway agencies, as documented in a synthesis of practice and from phone calls with maintenance practitioners. It concludes with the emerging practices identified from the literature and as part of the interviews with data collection vendors. As much as possible, the summary of practice focuses on the technology used for collecting, processing, and managing roadway asset data. A great deal of information is also available on assessing the condition and performance of roadway assets, but that information was considered to be outside the scope. However, similarities in the technology used for establishing an inventory and conducting a condition survey exist. For instance, cameras and other equipment can be added to the vans used for conducting pavement management surveys to facilitate the extraction of asset inventory data (AASHTO 2006). The information obtained through the investigation into current practices, including the interviews with state DOT practitioners, provided much of the basis for the information contained in the Guide. TECHNOLOGY BEING USED TO ESTABLISH ASSET INVENTORIES There are several different methodologies being used to collect inventory information and to assess the condition of roadway assets. These techniques range from manual surveys that use processes where people are directly involved in the observation or measurement of pavement surface properties without the benefit of automated equipment (McGhee 2004) to automated surveys that involve data collected by imaging or by the use of noncontact sensor equipment (McGhee 2004). Today s manual surveys often take advantage of hand-held computers and other forms of technology that have greatly improved the efficiency of data collection and processing activities. Data collected using automated methods can be evaluated using software tools that automate the extraction and interpretation of the data (commonly referred to as fully automated) or through semi-automated methods that require some human interaction to extract or interpret the data. Some agencies are also using mobile imaging with or without Light Detection and Ranging (LiDAR), a three-dimensional (3-D) technology that can rapidly acquire a great deal of highly-detailed geospatial information. Each approach has certain advantages and disadvantages, which may include some of the following (McGhee 2004): Manual data collection techniques are most appropriate for assets that are not readily available from the travel lanes. Traditionally, the methodology is slow and safety of the crews may be an issue, but the recent use of hand-held computers for recording survey information has increased the efficiency of this process. 4

Final Report Automated (or mobile) data collection techniques allow multiple assets to be assessed at the same time while traveling at traffic speeds. However, the assets must be visible from the travel lane and the equipment typically requires specialized equipment and operators. Automated processing allows large amounts of data to be available quickly, but the interpretation is constrained by the computer s ability to recognize certain types of assets and their characteristics. Semi-automated processing is slower than automated processing, but it provides for human interpretation of data from the field in a safe, workstation environment. Mobile LiDAR data can be collected quickly and with high accuracy for 3-D mapping, but the amount of data collected can require substantial resources to process. AVAILABLE REFERENCES ON BUILDING AN ASSET INVENTORY Data Collection Techniques Recognizing that an asset inventory is a key component to a comprehensive asset management program, AASHTO developed the Asset Management Data Collection Guide to address the data collection needs associated with asset management (AASHTO 2006). This reference documents the struggles transportation agencies have had to collect, store and analyze comprehensive inventory data for non-pavement and non-bridge assets and the advances that have occurred with handheld mobile computing devices in conjunction with Geographic Information Systems (GIS). The report also provides guidance on prioritizing the assets to include in the inventory based on asset category, rank, and relative importance to the agency (AASHTO 2006). Other factors, such as asset value, the availability of data collection protocols, the ease of evaluation, the overall value to users, and data collection frequency, are also factors to consider when prioritizing assets. For those items included in the inventory, the report outlines the necessary decisions to assess the condition of the asset, including the method of assessing performance, the level of detail and accuracy needed, inspection frequency, and sampling strategy. A separate study conducted for NCHRP investigated the use of asset management principles for managing ancillary assets other than pavements and bridges. Included in the report is a hierarchy intended to serve as the basis for classifying information on these assets, which includes asset class, asset elements, and sub-elements as appropriate (Rose et.al. 2014). The report also provides guidance for managing signs, traffic signals, markings, barrier systems, and lighting with information for establishing the inventory, assessing conditions, and estimating service life. The AASHTO Asset Management Data Collection Guide compares the advantages and disadvantages associated with manual, mobile and satellite data collection techniques (AASHTO 2006). For instance, manual data collection methods are reported to be relatively accurate and they allow access to assets that are not visible from the road; however, the process can be slow and labor intensive (AASHTO 2006). It also exposes agency personnel to safety hazards caused by interactions with traffic. The Guide identifies the collection of multiple data items at traffic speeds as an advantage to mobile data collection processes (AASHTO 2006). However, it is only suitable for assets that can be seen from the road and it requires special equipment that often forces agencies to contract out the data collection services. The suitability of different data collection methods for various types of assets was documented by Zimmerman and Stivers (2007) and is presented as Table 1. 5

Final Report Table 1. Suitability of different methods of data collection (Zimmerman and Stivers 2007). Asset Categories Asset Types Data Collection Method Asset Categories Asset Types Data Collection Method Drainage Culvert Manual Traffic Items Signal Manual Curb and gutter Manual Sign Manual or Mobile Sidewalk Manual Pavement markings Manual or Mobile Ditch Manual Pavement marker Mobile Drop inlet and storm drain Manual Overhead sign structure Manual or Mobile Erosion control Manual Traffic barrier/median barriers Manual Under or edge drain Manual Highway lighting Manual or Mobile Roadside Fence Manual or Mobile Guardrail & Attenuators Guardrail Manual or Mobile Grass mowing As Needed Guardrail end treatments Manual or Mobile Brush As Needed Impact attenuator Manual or Mobile Landscaping Manual Other Facilities Tunnels Manual Sound barrier Manual Rest areas Manual Pavement Shoulder Manual or Mobile Lane, paved Lane, unpaved Manual or Mobile Manual or Mobile Weigh stations Roadside Graffiti Roadside Litter Manual Manual Manual or Mobile The AASHTO Asset Management Data Collection Guide indicates that it might be cost-effective for state highway agencies using automated technology to collect pavement condition information to begin extracting asset inventory data for some assets from the images that are collected. It also identifies the manual method as the most commonly used method for establishing a roadway asset inventory and assessing asset condition (AASHTO 2006). The feasibility of using automated equipment for building roadway inventories was documented in an NCHRP report that describes the use of technology for georeferencing the data (NCHRP 2000). A 2004 NCHRP Synthesis describes the use of automated data collection devices in state highway agencies (McGhee 2004). A survey conducted for the synthesis found that the most commonly employed methods of automated data collection make use of acoustic or laser sensors, and image-processing tools. At that time, digital imaging was reportedly preferred over analog imaging techniques (McGhee 2004). In 2005, the FHWA conducted a case study to document the techniques being used by eight state highway agencies to manage roadway safety hardware, such as longitudinal barriers, crash cushions, attenuators, end treatments, breakaway supports, and work zone hardware (FHWA 6

Final Report 2005). The study found that the New Mexico Department of Transportation (DOT) used an automated vehicle to capture right-of-way images for its state-maintained roadway safety assets. The inventory is updated, and conditions assessed, by field personnel equipped with hand-held computers with Global Positioning Satellite (GPS) features. In the office, a Virtual Drive was set up to enable agency personnel to view highway segments and to determine the adequacy of signage, guardrail treatments, and other roadway hardware. The report indicates that the other seven states also use right-of-way images to build their asset inventories and handheld devices to collect condition information. The report concluded that the use of right-of-way imagery and GPS coordinates at a workstation was a common approach to establishing an asset inventory among state agencies and that manual data collection methods were more commonly used to collect condition information on these assets (FHWA 2005). The use of automated and manual data collection techniques for collecting information on roadway safety hardware was discussed with participants as part of a peer exchange on Asset Management and Safety. Meeting participants reported that most were using manual field inspections in combination with one or more additional data collection methods (FHWA 2011). For example, the report indicates that surveys for night retroreflectivity of signs could be done manually in conjunction with automated surveys featuring GPS capabilities to improve location accuracy. The feasibility of using LiDAR to inventory roadway assets was also discussed, but participants indicated that the cost-effectiveness of the technology had not yet been demonstrated. The Asset Management and Safety Peer Exchange participants reported that lack of resources had been an obstacle to having data available on all safety assets. They indicated that they often had robust inventories for signals, signs, guardrails, and lighting, but little information on road edge delineators, for example (FHWA 2011). They also discussed the level of detail required for inventorying and assessing the condition of safety assets and reported that they often had trouble effectively using all the data collected. The most pressing issues identified by the Peer Exchange participants concerning their safety asset inventories included (FHWA 2011): Location referencing accuracy and consistency. Temporal referencing accuracy and consistency. Availability of trained personnel. Availability of tools and systems in order to integrate safety-related asset data with other data. A Peer Exchange conducted in 2009 with state maintenance personnel indicates that manual data collection techniques for inventorying and assessing the condition of roadway assets are used most often, even though their agencies were using automated techniques for pavement distress surveys (Adams et al. 2009). The participants in the peer exchange expressed interest in some of the new technological advancements (e.g., LiDAR), but questioned the cost-effectiveness of the technology. A domestic scan that was conducted in October 2011 investigated best practices for collecting and reporting highway maintenance performance information (NCHRP 2012). The participants confirmed that most participants were using some type of manual survey to collect maintenance inventory and condition information. However, some of the participants reported that technology had advanced to the point that it could improve the efficiency of data collection 7

Final Report activities. For example, the Utah DOT reported on a pilot study they had conducted that showed that data collected using handheld devices as part of their manual surveys could be collected as quickly and as accurately as data collected with automated data collection vans (NCHRP 2012). In a study conducted for the North Carolina DOT, the Institute for Transportation Research Education at North Carolina State University compared the results of both manual and mobile data collection techniques to establish a roadway asset inventory (Cunningham et al. 2013). The results indicate the mobile data collection vehicles located roadway assets accurately, as long as there were no obstructions from landscaping or other vehicles (Cunningham et al. 2013). Mobile data collection methods were reported to show promise for accurately identifying feature characteristics, such as asset type, and measurements of asset height and road grade were measured within allowable tolerances. These devices were found to be less accurate with measurements parallel to the direction of traffic, such as offset distance or width. There were also several point features, such as drop inlets or attenuators, which proved to be difficult to georeference (Cunningham et al. 2013). An ongoing Strategic Highway Research Program 2 (SHRP2) study compared the accuracy of data produced by mobile imaging techniques with data collected from a manual inventory of eleven different roadway attributes. The initial findings indicate that there is a high degree of agreement between the two approaches in terms of the total number of items counted, but less accuracy in identifying the geospatial location of each item (Smadi 2014). LiDAR The use of LiDAR in transportation agencies was explored in a report prepared by the Wisconsin DOT (WI DOT 2010). This study documents applications for three types of LiDAR, including airborne, mobile, and terrestrial LiDAR. It explores applications in surveying, highway design, corridor development, critical infrastructure protection, traffic flow, highway safety, rock cuts, and geology. The study found that there are three technical aspects to LiDAR s use in these applications that are being refined: 1) data collection and analysis techniques, 2) error and accuracy measures, and 3) the integration of LiDAR and photogrammetry (WI DOT 2010). The report also includes a list of useful references on the use of LiDAR. The Michigan DOT also explored the use of remote technology to inventory highway roadside assets, comparing the use of aerial and mobile imaging with LiDAR, mobile imaging with photo logging, and manual data collection (MDOT 2014). The report claims that the use of aerial LiDAR eliminates worker exposure to traffic and represents the fastest mode of data collection. The output from the process is a point cloud, with millions of data points spatially located within a 3-D file. The aerial LiDAR equipment can be attached to a vehicle driven at traffic speeds or it can be carried on an aircraft flown at approximately 1,600 feet. The data collected using this technology can reportedly be collected once and used for a variety of applications where the height, width, and depth of an asset is needed. The disadvantages reported with aerial LiDAR indicate that the 900-foot perspective on the asset is better for statewide issues rather than project-specific issues, it is expensive to collect and process, and it is difficult to capture data in mowable areas. The Michigan DOT report describes mobile imaging as cameras set up on a vehicle to capture a variety of images at a regular interval, such as fifty feet, using panoramic and side-mounted cameras (MDOT 2014). The combination of images captures assets that can be seen from the roadway, such as signs, signals, and other roadside hardware. The images can be viewed at a 8

Final Report workstation, so a user can virtually drive any route captured by the cameras, without the safety hazard of being in the field or the time required to visit a remote site. The report indicates that some vendors have automated the processes to identify certain assets within the image set. Manual data collection is described in the report as an inspection that requires personnel in the field to record asset inventory information manually. The study found that manual methods are most appropriate for assets such as culverts, which are difficult to see from the travel lane (MDOT 2014). The Michigan DOT conducted a pilot project to collect data on twenty-seven predetermined assets. The results were analyzed in terms of cost-effectiveness and the amount of time required to collect and process the data. The report indicates that aerial LiDAR was twice as expensive as collecting the data manually, without much time savings (MDOT 2014). However, the researchers suggested that the high costs were due, in part, to the use of short segments in the pilot study. They hypothesize that the costs would have been reduced if the sections had been longer. Other findings from the study included the following (MDOT 2014): Remote technologies were able to collect data on most assets studied with the exception of those assets not readily visible form the roadway, such as culverts. LiDAR is appropriate for some applications, but was found to produce a level of detail that was not needed for the assets included in the study. Mobile imaging technology is an effective way to collect highway asset data at a reduced overall cost. It reportedly decreases worker exposure to traffic, speeds up data collection, and improves the accuracy of the data. NCHRP Report 748, Guidelines for the Use of Mobile LiDAR in Transportation Applications, provides a comprehensive summary on procurement considerations, data mining techniques, and quality control associated with the use of this technology. The report summarizes several advantages associated with the use of mobile LiDAR. For instance, inventory data can be collected for several assets in one pass and the data can be collected at highway speed without placing workers in traffic. This makes LiDAR very cost-effective in most situations. In addition, the results can be shared among different departments, so there is often wider use of the data collected. However, the report recognizes that while LiDAR is one of the tools in the toolbox that transportation agencies should consider, a benefit/cost analysis should be conducted to determine the cost-effectiveness of mobile LiDAR for specific applications (NCHRP 2007). Data Quality The importance of quality data is also addressed in the literature. For example, in 2013 the FHWA published its Practical Guide for Quality Management of Pavement Condition Data Collection. Although the focus of the Guide is on the collection of pavement condition data, it provides a useful framework for implementing quality management practices that can be used for a variety of data collection efforts, including measures related to resolution, accuracy, and repeatability; responsibilities for managing the quality of the data before, during, and after data collection; quality control processes; and quality acceptance processes (FHWA 2013). 9

Final Report The Utah DOT proactively addressed some of the data quality challenges it faced as it adapted its approach to managing high-priority roadway assets (such as guardrails, traffic signals, signs, drainage, and pavement markings) using new technology. For example, the agency initially focused its attention on the high-priority assets and developed guidelines to ensure consistency in inspection and data collection procedures. To aid in developing its comprehensive roadway asset inventory, and to train personnel in the use of the new system, peer forums were conducted to disseminate best practices (FHWA 2012). The Utah DOT has set a target of two years before it has complete asset inventories for all its maintenance assets. The agency recognizes that establishing an asset inventory takes time, but anticipates benefits in being able to better allocate budgets to address needs based on performance in the future (FHWA 2012). STATUS OF ASSET INVENTORIES IN STATE DOTS NCHRP recently published a synthesis of field inspection practices associated with Maintenance Quality Assurance (MQA) programs. The results of this synthesis provide a current snapshot of the status of asset inventories in the twenty-eight state DOTs with MQA programs in place that responded to the survey (NCHRP 2015). General observations from the synthesis include the following (NCHRP 2015): With the exclusion of pavements and bridges, the establishment of inventories for culverts, overhead sign structures, signs, signals, variable message boards, impact attenuators, pavement markings, guardrail end treatments, and rest areas has been completed or is in the process of being established in more than twenty of the twentyeight agencies that have formal Maintenance Quality Assurance (MQA) programs in place. By far, manual methods of data collection are most common for these assets. However, there are indications that the use of handheld computers and GPS units are being increasingly incorporated into the survey process. Additionally, some states are using automated mobile approaches to establish the inventory of some assets that are visible from the roadway. The use of automated equipment to build an asset inventory appears to be of interest for a number of states. The NCHRP Synthesis, Maintenance Quality Assurance Field Inspection Practices, asked respondents with MQA programs in place to complete a survey describing the status of their asset inventory and the methods used to assess the condition of the assets. For purposes of this report, information on the status of asset inventories and the methods of building the inventory are highlighted for the following asset categories and features: Drainage assets, including culverts, flumes, curbs and gutters, sidewalks, ditches or slopes, drop inlets, and underdrains/edgedrains. Roadside, including fence, landscaping, plant beds, and sound barriers. Pavement, including paved shoulders, unpaved shoulders, and paved roadways. Bridge, including all bridge structures greater than 20 feet in length. Traffic items, including signals, signs, pavement markings, pavement markers, guardrail end treatments, overhead sign structures, impact attenuators, and protective barriers. 10

Final Report Special facilities, including rest areas, tunnels, weigh stations, and traffic monitoring systems. The inventory status of each asset category is presented separately. Drainage Assets According to the information presented in Figure 1, few states have developed complete inventories for all of the elements included in this category. Of the seven elements included in the survey, the largest number of states have either established or are in the process of establishing an inventory of their culverts. More than half of the agencies responding to the survey have established, or are in the process of establishing, inventories for curb and gutter, drop inlets, ditches or slopes, and sidewalks. A smaller number of states indicate that they have established inventories for flumes or underdrains and edgedrains. Number of Responses Roadside Assets Figure 1. Inventory status of drainage assets (NCHRP 2015). Sound barriers, fences, landscaping, plant beds are all examples of roadside assets. As shown in Figure 2, half of the state agencies that responded to the survey have established, or are in the process of establishing, asset inventories for sound barriers, but the status of inventories for other roadside assets are not as far along. 11

Final Report Number of Responses Figure 2. Inventory status of roadside assets (NCHRP 2015). Pavement and Bridge Assets The asset inventories for pavements and bridges in state DOTs are essentially complete. The inventories of other assets in this category, including paved and unpaved shoulders are shown in Figure 3. Number of Responses Figure 3. Inventory status of pavement assets (NCHRP 2015). Traffic Assets The traffic asset category includes a variety of safety-related assets, such as signs, signals, pavement markings and markers, guardrail end treatments, overhead sign structures, and variable message boards. The status of asset inventories in the twenty-eight state DOTs with MQA programs in place is shown in Figure 4. As shown, there are at least three assets in this category 12

Final Report in which more than half of the states have established complete inventories (i.e., overhead sign structures, signals, and variable message boards). At least half of the agencies responding to the survey have begun creating the asset inventory for each of the assets listed. Number of Responses Figure 4. Inventory status of traffic assets (NCHRP 2015). Special Facilities Some state DOTs are responsible for the maintenance and management of special features, such as rest areas, tunnels, weigh stations, and traffic monitoring systems. As shown in Figure 5, these inventories are fairly well established in agencies that manage these types of assets. Number of Responses Figure 5. Inventory status of special facilities (NCHRP 2015). 13

Final Report EMERGING TRENDS As technology improves and becomes more commonly available to practitioners, it is likely that more of the data collection, processing, and managing functions will become more automated than in the past. Additionally, as practitioners become more familiar with the technology, they will find new and innovative applications that expand its use within the agency and improve its cost-effectiveness. In addition to changes in technology, transportation agencies are continually responding to legislative, funding, and system demand changes that impact the way they do business. The increased use of public-private partnerships and performance-based warranty contracts in transportation agencies are examples of agency responses to the changing operational environment. As a result, it is important that transportation agencies develop business processes that provide enough flexibility to be able to respond to the changes they face. It is also important that transportation agencies realize that it takes time for changes, such as new technology, to be incorporated into the on-going business activities. The Utah DOT, for example, indicates that it generally takes between two and four years for applications of new technology to mature and become integrated into routine work activities (FHWA 2012). Based on the experience of the research team, and the knowledge gained during this project, the following general trends are observed in how state DOTs are building or updating their roadway asset inventories. As agencies realize the benefits associated with performance-based decision making, there will be increasing interest in using performance data on assets other than pavements and bridges for establishing and defending budget needs, allocating funds, and documenting the effectiveness of investments. Additionally, as agencies evaluate and manage risks from an enterprise level, they will recognize that the availability of reliable data on key assets can help mitigate some of the agency s risks. As a result, transportation agencies are expanding the scope of their data collection projects and will continue adding assets to their roadway inventories. Limited resources will continue to force transportation agencies to be more effective with their data collection efforts by finding additional uses for the data and/or collecting more data with each pass. The development and use of data governance standards and the availability of tools that allow the integration of data sets, will become increasingly important to realize these efficiencies. Transportation agencies are recognizing that their workforces need to develop new skills to fully utilize the new technology or that non-traditional hires are needed to provide the necessary capabilities. Individuals with GIS training, database skills, communication experience, and working knowledge of computers and technology are helpful in turning data into useful knowledge. As traditional DOT functions are privatized, transportation agencies will need to develop strategies for using data to monitor the contractor s activities. For instance, pavement condition data used to monitor a treatment warranty may require a higher level of data reliability than the data used to report network conditions. How to meet these demands using technology will become an important priority in the future. 14

Final Report EMERGING TECHNOLOGY As technology continues to advance, a number of technologies that are currently either in the development phase or undergoing feasibility testing may become viable additions to the methodologies considered in the Guide. Some of the more promising technologies available are presented in the remainder of this section. 360-Degree Camera One emerging technology is the development of a camera that consists of six lenses, positioned complementary to one another, providing a 360-degree horizontal perspective of an area. The camera uses mathematical algorithms to stitch together the various images to create the 360- degree view of the roadway. One of the six cameras is positioned vertically rather than horizontally to create a spherical view. These images, as in photogrammetry, are linked with GPS coordinates to identify field locations of the extracted data. This technology may provide additional benefits to traditional photogrammetry techniques in situations that would benefit from a 360-degree perspective, such as intersections and interchanges. Flash LiDAR In contrast to mobile LiDAR in which every single point is illuminated individually with a laser, flash LiDAR illuminates a whole scene at once. As a result, each pixel provides an indication of the amount of time that passed for the camera s laser flash pulse to hit the targeted asset and bounce back to the camera s focal plane. The time measurements are resolved using the speed of light, resulting in a 3-D image from the depth measurements for each point. Flash LiDAR is currently being tested for applications in the military and automobile industry due to its ability to provide real-time information. Flash LiDAR is also referred to as time-of-flight (TOF) cameras. Airborne LiDAR Aerial or airborne LiDAR has been around for several years but its use has been limited due to Federal Aviation Agency (FAA) flight restrictions that were imposed to avoid any conflicts with air traffic. Airborne LiDAR captures data on a scale that lends itself more to design and planning activities rather than building roadway asset inventories. For example, scans from airborne LiDAR have been used to create 3-D models of complex objects, such as piping networks, roadways, archeological sites, buildings, and bridges. Airborne LiDAR has been used on large, civil engineering projects to assist with grading, utilities, drainage analysis, erosion control, and roadway design. It has also been used by the military and in the archaeological and agricultural fields. There are several advantages to airborne LiDAR that make the technology appealing to the transportation community. For example, objects can be measured remotely without interfering with traffic. Additionally, the equipment can be operated under a variety of weather conditions and its sensors are not affected by low sun angles. Airborne LiDAR can even be used at night. In the past year, the FAA has granted approval to four companies to fly commercial drones to conduct aerial surveys, monitor construction sites, and inspect oil flare stacks (USA Today 2014). The results of the trials are expected to influence the future use of this technology. 15

Final Report Driverless Cars Automobile manufacturers have increasingly shown interest in the concept of driverless cars and the everyday use of this technology could soon become a reality. If that were to be the case, transportation agencies may have to shift their data collection priorities since driverless cars could require different roadway features to operate effectively. For example, striping could become increasingly important to keep the cars in the driving lane. Transportation agencies will face a major paradigm shift in terms of data collection and asset performance as this new technology becomes more common. 16

Final Report CHAPTER 3 CONCLUSIONS The availability of roadway asset inventory information is a fundamental component of a comprehensive asset management program that uses performance data to drive investment decisions. Although many transportation agencies have complete inventories of their pavements and bridges, fewer agencies have complete inventories for the other roadway assets that they manage and maintain. As a result, transportation agencies have been limited in their ability to determine maintenance needs and to convey these needs to their stakeholders. Today, data collection and processing technology has advanced to the point that it can improve the cost-effectiveness of establishing an asset inventory. Some agencies have found that by consolidating disparate data collection efforts, several data needs can be satisfied quickly without impacting traffic. However, since no single technology is appropriate for all applications, guidance was needed to help agencies evaluate the appropriateness of the different options for different situations. This report documents the state of practice and the information obtained through the project tasks led to the development of the Guide, which is included as an Attachment to this report. The investigation into the approaches used to establish roadway asset inventories indicate that the manual methods of collecting the data, often supplemented with handheld technology, remain the most common approaches in state DOTs. Some agencies that are using automated technology for conducting pavement management surveys have used the same tools to extract some asset information. At least one state is using LiDAR in an attempt to inventory all of its roadway assets within the next several years. The information provided demonstrates the feasibility of using automated technology for establishing or updating an asset inventory. The differentiation between using photogrammetric methods versus remote sensing technology (such as LiDAR) depends on the specific needs of the agency. In general, LiDAR is advantageous in situations where vertical clearances or highly accurate offset distances are needed, or if the data can be used to support other agency functions, such as planning and/or design. Even though a technology is feasible, there are many additional considerations that have to be taken into account when selecting a technology for establishing a roadway asset inventory. These considerations include resources available, asset location (i.e., visibility from the road), number of assets to be included, level of detail required, and safety requirements. Additionally, each methodology has different requirements in terms of equipment needed, level of technical expertise required to operate the equipment, and data storage requirements that should be taken into account. These considerations are all addressed further in the Guide that was developed as a result of this project to serve as a useful resource to transportation agencies interested in establishing or expanding their roadway asset inventory. FUTURE RESEARCH NEEDS The results of this research identified several gaps in current knowledge that could benefit from additional research. Recommendations for further research include the following: Strategies for cost-effectively expanding the application and use of remote sensing technology within a transportation agency. 17

Final Report Guidance with establishing quality measures for roadway asset inventory items using automated data collection processes. Methods of processing automated data to reduce the time between data collection and data delivery. Identification of successful strategies used by transportation organizations to facilitate the timely implementation of innovations into agency practices. 18

Final Report REFERENCES Adams, T., S. Janiowiak, W. Sierzchula, and J. Bittner. 2009. Maintenance Quality Assurance Peer Exchange 2. Project 08-15. Midwest Regional University Transportation Center, University of Wisconsin, Madison, WI. American Association of State Highway and Transportation Officials (AASHTO). 2006. Asset Management Data Collection Guide. Task Force 45 Report. American Association of State Highway and Transportation Officials, Washington, DC. Cunningham, C. M., D. J. Findley, K. Hovey, P. B. Foley, J. Smith, T. Fowler, J. Chang, and J. E. Hummer. 2013. Comparison of Mobile Asset Data Collection Vehicles to Manual Collection Methods. North Carolina Department of Transportation, Raleigh, NC. Federal Highway Administration (FHWA). 2005. Roadway Safety Hardware Asset Management Systems Case Study. FHWA-HRT-05-073. Federal Highway Administration, Washington, DC. Federal Highway Administration (FHWA). 2011. Asset Management and Safety Peer Exchange Report. FHWA-HRT-12-005. Federal Highway Administration, Washington, DC. Federal Highway Administration (FHWA). 2012. Managing and Maintaining Roadway Assets: The Utah Journey. A Transportation Asset Management Case Study. Federal Highway Administration, Washington, DC. Federal Highway Administration (FHWA). 2013. Practical Guide for Quality Management of Pavement Condition Data Collection. Federal Highway Administration, Washington, DC. McGhee, K. H. 2004. Automated Pavement Distress Collection Techniques. NCHRP Synthesis 334. National Cooperative Highway Research Program, Transportation Research Board, Washington, DC. Michigan Department of Transportation. 2014. Monitoring Highway Assets with Remote Technology. RC -1607. Michigan Department of Transportation, Research Administration. Lansing, MI. National Cooperative Highway Research Program (NCHRP). 2000. Collection and Preservation of Roadway Inventory Data. NCHRP Report 437. Transportation Research Board, National Research Council, Washington, DC. National Cooperative Highway Research Program (NCHRP). 2007. Use of Mobile LiDAR in Transportation Applications. NCHRP Report 748. Transportation Research Board, National Research Council, Washington, DC. National Cooperative Highway Research Program (NCHRP). 2012. Best Practices in Performance Measurement for Highway Maintenance and Preservation. Scan Team Report, NCHRP Project 20-68A, Scan 10-03. Transportation Research Board, National Research Council, Washington, DC. 19

Final Report National Cooperative Highway Research Program (NCHRP). 2015. Maintenance Quality Assurance Field Inspection Practices. NCHRP Synthesis 470. National Cooperative Highway Research Program, Transportation Research Board, Washington, DC. Rose, D., K. Shah, J. P. O Har, and W. Grenke. 2014. Transportation Asset Management for Ancillary Assets. NCHRP 08-36, Task 114 Final Report. National Cooperative Highway Research Program, Transportation Research Board, National Research Council, Washington, DC. Smadi, O. 2014. Personal notes transcribed from an interview. USA Today. 2014. FAA Lets Four Companies Fly Commercial Drones. USA Today. Accessed December 10, 2014: http://www.usatoday.com/story/money/business/2014/12/10/faadrones-trimble-vdos-clayco-woolpert-amazon/20187761/ Wisconsin Department of Transportation (WI DOT). 2010. LiDAR Applications for Transportation Agencies. Transportation Synthesis Report. Wisconsin Department of Transportation, Madison, WI. Zimmerman, K. A. and M. L. Stivers. 2007. A Guide to Maintenance Condition Assessment Systems. NCHRP Project No. 20-07, Task 206. National Cooperative Highway Research Program, Transportation Research Board, Washington, DC. 20

NCHRP Project 20 07/Task 357 A GUIDE TO COLLECTING, PROCESSING, AND MANAGING ROADWAY ASSET INVENTORY DATA FINAL VERSION June 2015 The information contained in this report was prepared as part of NCHRP Project 20-07, Task 357, National Cooperative Highway Research Program. SPECIAL NOTE: This report IS NOT an official publication of the National Cooperative Highway Research Program, Transportation Research Board, National Research Council, or The National Academies.

TABLE OF CONTENTS CHAPTER 1 INTRODUCTION... 1 PURPOSE OF THIS GUIDE... 1 GUIDE ORGANIZATION... 1 GUIDE FOCUS... 2 USING THE GUIDE... 2 CHAPTER 2 DATA COLLECTION METHODS... 3 MANUAL TECHNIQUES... 3 AUTOMATED TECHNIQUES... 7 Photogrammetry... 7 Mobile LiDAR... 10 SUMMARY... 12 ADDITIONAL READING MATERIAL... 12 CHAPTER 3 GUIDELINES... 14 STEP 1: GETTING READY TO SELECT A METHODOLOGY... 15 Select Assets to Include in the Inventory... 15 Determine Resource and Other Constraints... 17 Identify Users... 17 Establish a Data Dictionary... 18 STEP 2: SELECTING A METHODOLOGY... 19 Evaluate Asset Visibility from the Road... 19 Consider Accuracy Requirements... 19 Assess Agency Maturity... 20 Consider Safety Requirements... 20 Evaluate Resources... 20 Identify Other Data Collection Efforts... 21 Summary... 22 STEP 3: COLLECTING THE DATA... 22 Secure Data Collection Equipment and/or Vendor... 22 Develop Data Collection Protocol... 23 Conduct Personnel Training and Equipment Calibration... 24 Conduct Quality Control and Acceptance Testing... 25 STEP 4: PROCESSING AND MANAGING THE DATA... 25 Develop In-House Technical Expertise... 25 Formulate Data Processing Procedures... 25 Provide Access to Data... 26 Address Organizational Issues... 27 Implement Data Governance Standards... 27 Develop Plans for Inventory Updates... 27 Other Considerations... 28 EXAMPLE... 29 The Scenario... 29 The Process... 29 ACCELERATING THE LEARNING CURVE... 30 Challenges and Possible Remedies... 31 Benefits Realized... 31 ADDITIONAL READING MATERIAL... 32 i

CHAPTER 4 FUTURE DIRECTIONS... 33 FUTURE MODIFICATIONS TO THE DATA COLLECTION PROCESS... 33 EMERGING TECHNOLOGIES... 34 360 Degree Camera... 34 Flash LiDAR... 34 Airborne LiDAR... 34 ADVANCEMENTS IN DATA PROCESSING TECHNIQUES... 35 SUMMARY... 35 ADDITONAL READING MATERIAL... 36 REFERENCES... 37 APPENDIX A SAMPLE DATA DICTIONARY...A-1 APPENDIX B TYPICAL CONTENT IN A DATA COLLECTION RFP...B-1 ii

LIST OF FIGURES Figure 1. Characteristics associated with manual data collection techniques....4 Figure 2. Extract from a manual data collection form used by the Alabama DOT (http://www.dot.state.al.us/maweb/frm/aldot%20condition%20assessment%20d ata%20collection%20form.pdf)....5 Figure 3. Characteristics associated with photogrammetry....9 Figure 4. Characteristics associated with mobile LiDAR....11 Figure 5. Guidelines for developing or updating a roadway asset inventory....16 Figure 6: Relation between decision making levels and detail and amount of data required (Flintsch 2006)....18 Figure 7: Relative comparison of resource requirements, data utility, and costs (not to scale)....21 Figure 8. Factors in selecting a methodology for building a roadway asset inventory....22 Figure 9. Screenshot of New Mexico RFI spatial map with assets identified (Hensing and Rowshan 2005)....27 LIST OF TABLES Table 1. Applicability of each data collection methodology to inventory roadway assets....13 iii

PURPOSE OF THIS GUIDE CHAPTER 1 INTRODUCTION The management, preservation, and improvement of the highway system is a critical component to the nation s economy. Transportation agencies are responsible for the maintenance and management of the highways, roads, bridges, and other physical assets that keep the public moving safely and reliably. To help make investment decisions for preserving these valuable assets, many transportation agencies collect inventory and performance data. While this information has been collected on roads and bridges for many years, constrained resources have kept many transportation agencies from collecting data on all of the other assets they maintain, including guardrails, culverts, and signs. However, with improvements in technology and in data management over the last several years, new methods of collecting and processing inventory data are being used by some agencies. This Guide, which was developed under National Cooperative Highway Research Program (NCHRP) Project 20-07, Task 257, serves as a resource to help transportation agencies make informed decisions on the type of methodology most appropriate for collecting asset inventory information and the considerations that must be taken into account for processing and managing the data. Although the study considered a variety of different types of methodology for building inventories, the Guide focuses on the three most commonly used approaches in today s transportation agencies: manual surveys and two forms of automated surveys (photogrammetric methods, which is also known as mobile imagery, and mobile LiDAR, which stands for Light Detection and Ranging). GUIDE ORGANIZATION The Guide is organized into three sections. The first section, Chapter 2, introduces the three methodologies that are commonly being used in transportation agencies for collecting asset information. The guidance is presented in Chapter 3, which outlines a 4-step process that involves the following activities: Step 1: Getting ready to select a methodology. Step 2: Selecting a methodology. Step 3: Collecting the data. Step 4: Processing and managing the data. Chapter 4 includes a summary of how the technology is expected to evolve in the next several years and how that will impact the decisions transportation agencies make today. To assist the reader in various aspects of collecting, processing, and managing the asset inventory data, additional information is provided as appendices to the guide. Included in the appendices is an excerpt from a data dictionary used to describe the characteristics of interest for the inventory (see Appendix A), and a summary of the content typically included in Requests for Proposals (RFP) if an automated data collection vendor is to be used (see Appendix B). 1

GUIDE FOCUS The Guide focuses on the collection, processing, and management of data used to develop and maintain a roadway asset inventory. Since most transportation agencies have complete inventories of their pavements and bridges in place, the Guide concentrates primarily on the other roadway assets maintained by state DOTs, such as guardrails, tower lighting, signs, and drainage features. The Guide covers only topics related to establishing and maintaining a roadway asset inventory. As a result, it does not address the development and use of performance criteria to monitor the level of service being provided to the traveling public or the funding needed to address maintenance needs. USING THE GUIDE This Guide was developed primarily for the maintenance personnel in state DOTs who are responsible for developing and maintaining a roadway asset inventory. It is designed to assist these individuals in determining the type of methodology most appropriate for building and maintaining the inventory, the technical and organizational considerations that should be addressed prior to building the inventory, and the data collecting and management issues that should be addressed with each of the three different approaches. The considerations described in the Guide are not unique to practices in state DOTs. The guidance provided in this document can be equally useful to asset and maintenance management personnel in cities, counties, or other transportation agencies. In addition to maintenance personnel, other practitioners may benefit from the information provided in this Guide. For instance, the information may help an agency that is using automated equipment for pavement management data collection find new uses for the digital images that are being collected. Similarly, an agency that is using a vehicle equipped with LiDAR for collecting inventory information may discover new applications for the technology to support the agency s design activities. 2

CHAPTER 2 DATA COLLECTION METHODS Data is a key element to making decisions for the maintenance and operation of roadway assets. Methods of collecting, analyzing, and reporting data have changed over the years as technology has advanced, especially for assessing highway pavement conditions using automated technology. The same equipment that is being used to collect pavement condition data can also be used, with just minor adjustments, to support other data collection activities within a transportation agency, including the establishment or update of roadway asset inventories. An agency s ability to make sound, defensible investment decisions relies in part on the availability of a comprehensive asset inventory, a method of assessing current conditions and performance, and tools for evaluating the impacts of different investment strategies on network performance. Establishing an inventory is a fundamental step in establishing a strong asset management program. The guidance provided in the next chapter concentrates on the use of three different types of technology, including manual techniques and two different types of automated technology. These techniques include the following: Manual techniques, which involve recording inventory information while walking or viewing assets from the windshield of a vehicle. Manual techniques may involve nothing more sophisticated than recording information with pencil and paper, or they may utilize Global Positioning Satellite (GPS) technology to locate assets in the field and/or handheld computers to record the field data. Automated techniques, which usually involve driving a specially-equipped vehicle at near-traffic speeds over the highway. The technology used by these vehicles, which is generally characterized as photogrammetry, typically includes laser sensors to monitor pavement-related characteristics (such as rutting and roughness) and digital cameras strategically placed to capture different roadway features. Some of the vans are also equipped with remote-sensing technology that measures distances by analyzing the reflected light from an asset after being lit by a laser. This additional technology is commonly referred to as mobile LiDAR. Each of these techniques is explored further in the remainder of this chapter, including a summary of the types of data that can be collected, the special characteristics that make the technique most appealing, and practical limitations that should be considered. MANUAL TECHNIQUES Manual techniques for building a roadway asset inventory typically involve surveys conducted by field personnel while either inside or outside of a vehicle. If the surveys are being conducted from the windshield of a vehicle, the surveys are typically done at a slow speed or from the shoulder of the road. Manual surveys can be very low-tech, meaning that little technology is used beyond pen and paper to record information, or they may make use of GPS technology and/or handheld computers to help automate parts of the process. Figure 1 shows manual data collection characteristics. According to a recent synthesis of practice, manual techniques are currently the most common methodology being used to build asset inventories for assets other than pavements and bridges (NCHRP 2015). When manual processes are used for building an inventory, the field crews often assess the condition of the assets at the same time so another trip 3

Data Collection and Processing Data are collected by personnel walking in the field or individuals recording information while looking out the windshield of a vehicle. Data is recorded either on paper or with hand-held computers. The information collected during a manual survey can be very detailed if necessary. Data has to be processed manually if collected on paper. If data are collected using hand-held computers, the information is uploaded into a software program prior to its use. Little to no prior technical expertise is required for data collection and processing activities. Provides a way to survey assets that are not visible from the roadway, such as drainage assets. This technique is the most common approach used by state highway agencies for developing an inventory and it is the only method being used to inventory drainage assets (NCHRP 2015). Application Can be used to inventory all roadway assets including ones that are not visible from the roadway. Many agencies use hand-held computers to enhance their data collection and processing efficiency. Value Added Does not require specialized equipment so it can be implemented easily at a relatively low cost. Assets that are not easily visible from the road can be surveyed. The physical presence of an inspector in the field usually leads to good quality data. Limitations The accuracy of distance measuring devices is limited to a few feet. Quality assurance and lack of consistency can be an issue if raters are not trained. Data collection is slower than other available methods and may expose raters to traffic. Figure 1. Characteristics associated with manual data collection techniques. 4

to the field is not required. An example of the type of data collection form often used for manual surveys is provided in Figure 2. The example was extracted from the full data collection form available from the Alabama Department of Transportation website (http://www.dot.state.al.us/maweb/frm/aldot%20condition%20assessment%20data%20coll ection%20form.pdf). Figure 2. Extract from a manual data collection form used by the Alabama DOT While the use of manual data collection forms is common, there has been a rise in the use of hand-held computers for entering data in the field. The development of software programs and mobile applications for the use in tablets and smart phones has made the data collection process more efficient and reliable than traditional methods. For instance, the use of tablets with cameras and GPS capabilities has not only allowed agencies to create more extensive inventories with GIS referencing, but has also reduced the amount of equipment needed in the field. These applications can be used off-line but also offer the option to instantly update a database located on a server when online. The Iowa DOT used this technology for creating a culvert and guardrail inventory and the Massachusetts DOT use it for improving and updating its ADA ramp inventory. Other agencies have used this technology for creating sign inventories. 5

Scope: Since manual techniques provide an opportunity for an inspector to walk to the location of most assets, an asset inventory can be established for virtually any roadway asset using this technique. For some assets, such as those that are not directly visible from the roadway surface (e.g., drainage assets), this is essentially the only viable data collection method available. Other techniques are limited in their ability to assess partially submerged assets or those located outside a driver s line-of-sight. With the increase use of hand-held computers and the development of software programs and mobile applications for collecting and reporting inventory information, the efficiency of manual processes has improved over traditional methods featuring paper forms. Data Collection: The inventory of roadway assets is established by driving along the roadway and stopping at the location of each asset to collect the required data. The data collected from this activity can be very detailed if necessary, but it does not have to be. Manual surveys do not require much technical expertise beyond training in what features and characteristics are being collected. Information collected in the field is usually recorded with paper and pencil, although the use of hand-held computers is becoming more popular as a way to improve efficiency and reduce the inconveniences associated with the use of paper, such as loss of data or delays in entering data. The use of handheld or tripod-mounted mobile laser distance measuring instruments has become a practical way for survey personnel to measure vertical clearances in the field, which has augmented the type of data that can be collected manually. The degree of reproducibility of the data with the manual method can be relatively low compared to other techniques if there is a lot of subjectivity to the process of collecting the data. This subjectivity can be reduced by regular training and certification programs for surveyors. Data Processing: If the data is collected using paper and pencil, the data processing will have to be done manually, which can be a very time-consuming process. It also introduces the possibility of errors in data entry. On the other hand, if hand-held computers are used to collect data, the data can be processed and summarized efficiently using the software provided with the data collection tool. Many applications now allow real-time entry of field data into a database. Applications: This technique provides the best option for collecting data for assets not directly visible from the roadway surface. Drainage assets, rest areas, and weigh stations are examples of the types of assets that are best surveyed using this methodology. The use of hand-held computers and mobile lasers has expanded the scope of this methodology and improved its efficiency. Advantages: The data collected using this method can range from being very detailed to very general. The physical presence of personnel inspecting each asset positively influences the type and extent of information that can be collected, which may lead to high-quality information. Since this method of data collection does not require specialized equipment, agency personnel can be trained to conduct the inspections and costs are limited to labor expenses. Overall, its biggest advantage is that it enables data to be collected on assets that are not easily viewed from the road. Limitations: The pace at which data is collected is slower than the other methods available. The significant involvement of personnel in the field could lead to safety issues since personnel must interact with traffic. It also introduces the possibility of human error or subjectivity in the data collection process. In addition, location measurements are made using distance measuring instruments (DMIs) and GPS units for georeferencing, which are only accurate within a few feet. 6

Also, in order to perform quality checks on the data collected, additional personnel must go out in the field, which adds to the cost. In summary, while manual surveys are limited by a comparatively slower rate of data collection than the other available options, this technique offers access to assets not directly visible from the roadway and requires little to no specialized equipment. This technique is very commonly used by maintenance personnel since surveys can be done by field personnel as they have time available. The work can also be contracted out, depending on the resources available to the agency. AUTOMATED TECHNIQUES There are two automated data collection techniques that are used for establishing roadway asset inventories: photogrammetry and mobile LiDAR. Both of these techniques are similar in that they use specially-equipped vehicles to collect the information at near traffic speeds. The processing of the data collected can be done using either automated or semi-automated techniques. In general, information collected using lasers is processed through automated techniques. Information to be extracted from the digital images may be processed using automated techniques, or the process could be semi-automated, meaning that the data are interpreted at workstations by individuals viewing the images. Photogrammetry Photogrammetry refers to the process of determining measurements from photographs or digital images, such as locating the position of a sign in the field (see Figure 3). The technique originally dates back to the mid-nineteenth century and it is still being used to create maps, drawings, and 3-D models. Many maps used today are created with photogrammetry, using photographs taken from aircraft. Photogrammetry has been adapted for use in conducting pavement condition surveys for more than 20 years. These surveys are conducted while a van outfitted with multiple cameras and other devices drives down a road at traffic speeds. The cameras are normally oriented so one of them is positioned to capture the road right-of-way (ROW) and several others are positioned in a downward-facing position to capture pavement surface details. Additional cameras can be added to the van and positioned at various angles to capture roadway features, such as signs, guardrails, and lighting structures. Scope: Photogrammetry is used by a number of transportation agencies across the United States to assess pavement conditions. With the addition of strategically-placed cameras on the van, this equipment can also be used to establish an inventory of roadside assets that are visible from the road. Data Collection: The roadway asset inventory data is obtained by driving along the roadway at highway speeds, with cameras mounted on a van. The camera placement influences the maximum field of view that can be captured in the image and any assets outside of the cameras view cannot be surveyed using this approach. GPS instruments added to the survey vehicle provide the geo-referenced coordinates that are needed to synchronize the data from the various cameras. The operation of the equipment requires a certain amount of technical expertise and training, but the equipment provides a relatively high degree of reproducibility. The quality of 7

the images are influenced by the quality of the cameras, the survey conditions, lighting conditions, and other factors. Data Processing: The asset characteristics are extracted from the digital images from the ROW or other cameras. Data extraction is most often conducted by personnel at a computer workstation, where the images can be viewed and specialized software provided by the data collection vendors can be used to calculate distances and track assets. Some vendors have developed tools that automate the data extraction process. The data provided by the automated processes can be viewed and verified at a workstation as part of a quality control process. Applications: This technique is a practical method for building a roadway asset inventory for assets that can be seen from the roadway. Supplemented with a manual survey for assets that are not visible from the road, photogrammetry provides a reliable method of collecting data on guardrails, sound barriers, retaining structures, fences, and other assets where a high degree of accuracy is not needed. The data extraction can be conducted at any time, even years after the surveys were conducted. Advantages: This method requires little human intervention during the data collection process, so consistency in the data is typically high. Data can be collected at highway speeds without requiring personnel or slow-moving vehicles to interact with traffic. Quality control checks of the data can be conducted at a workstation, which eliminates the need for follow-up surveys in the field. The technique is very cost-effective, especially if used for multiple applications, such as conducting pavement management surveys and building asset inventories. Limitations: Only assets visible from the roadway surface can be inventoried using this method. The amount of detail that can be collected is limited to what can be viewed on the camera image. Measurement accuracy typically falls within a few feet. It is limited by the capabilities of the DMIs and GPS units used for georeferencing. Photogrammetry has traditionally not been used for certain data elements, such as vertical clearances, but mobile laser distance measuring devices can be added to overcome this limitation. In summary, photogrammetry is a viable option for building an asset inventory for many roadway assets. It is most economical when ROW cameras are combined with lasers and downward-facing cameras used for pavement management surveys. Data can be collected at highway speeds at accuracies within a few feet without requiring personnel to interact with traffic. Additionally, this methodology makes quality checks easy to complete because of the ability to quickly review images at a workstation. While the use of this method is limited to assets visible from the roadway, it can be combined with manual surveys to take advantage of the benefits associated with each methodology. 8

Data Collection and Processing The data can be collected at the same time as data for other applications by adding a ROW camera and other speciallypositioned cameras. The surveys are conducted at traffic speeds. The asset inventory is created by extracting data using a data processing interface that is typically licensed from the equipment manufacturer. There is limited specialized knowledge required to build an inventory using this methodology. Data can be extracted from images at a time that is convenient to the agency. For instance, the data can be extracted a year after the surveys were conducted if resources are not available prior to that. Quality checks on the data can be conducted at the workstation without requiring personnel to go out to the field. Application Can be used to inventory all assets visible from the roadway. \ Data can be collected for multiple assets simultaneously at highway speeds. Value Added Increases consistency by limiting the amount of human intervention. Images can be reused for multiple purposes without another survey. Improves safety by reducing the number of personnel in the field. Limitations The accuracy of the roadway asset location may be limited to a few feet. Dimensions of assets cannot be extracted accurately. Assets that are not visible from the roadway surface cannot be inventoried using this method. Figure 3. Characteristics associated with photogrammetry. 9

Mobile LiDAR Mobile LiDAR is a remote sensing technology that measures distance by illuminating a target with a laser and analyzing the reflected light (see Figure 4). Mobile LiDAR is most commonly used to make high-resolution maps, with applications in areas such as archaeology, geography, and geology. Its application and use for asset management applications has been growing in recent years. Since many agencies do not have experience using mobile LiDAR, NCHRP produced a report outlining guidelines for using mobile LiDAR in transportation applications (NCHRP 2007). Scope: Mobile LiDAR can locate objects in the field to a high level of precision, within 3 in. up to a range of approximately 250 feet. LiDAR produces a 3-D point cloud that can be used to develop offsets or to measure vertical clearances. Similar to photogrammetry, this method is limited to assets directly visible from the roadway within the range of the camera. Data Collection: Mobile LiDAR is a 3-D measurement technology that can rapidly acquire a substantial amount of highly-detailed geospatial information. Additional sensors, such as cameras, reflectometers, laser crack measurement systems, or inertial profilers can be mounted on the same vehicle to collect additional information at the same time as the mobile LIDAR data acquisition. The data are collected while traveling at highway speeds but are limited to assets visible from the roadway. The use of mobile LiDAR requires a significant amount of technical expertise associated with the processing and use of the data. The point clouds generated by mobile LiDAR result in large files so data storage can be an issue. The technique has a high degree of reproducibility that is most influenced by the distance from the source and line of sight. Data Processing: Once the data is collected, some amount of processing is necessary for the data to be georeferenced. Roadway inventory assets are extracted automatically from a point cloud using proprietary software, usually developed by the data collection equipment manufacturer. Data extraction can be conducted at any time, even years after the surveys were conducted. Applications: This method can be used in conjunction with a manual survey to capture assets that are not visible from the road. The equipment is used effectively to measure asset features, such as pavement widths, that require a relatively high degree of accuracy. The equipment can be used to determine vertical clearances and to capture information about tunnels. Advantages: This technique requires little human intervention during the data collection process so it provides a consistent approach to collecting roadway inventory data. The use of mobile LiDAR improves safety by eliminating the need for personnel and slow-moving vehicles to interact with traffic. Data quality control checks can be conducted by reviewing the data at a workstation without requiring a new survey or forcing field personnel to drive back to the site of the asset for confirmation. Mobile LiDAR vehicles collect data at traffic speeds and provide a high degree of accuracy (± 3 in.). This method is also effective for estimating vertical measurements, such as vertical clearances. 10

Data Collection and Processing LIDAR is a remote sensing technology that collects data in 3-D point clouds. This technology is being used for a wide range of applications outside of asset management. Data is captured by a van driven at traffic speeds with a LiDAR sensor mounted on top and paired with a scanner, photo detector, and GPS. The equipment can be coupled with lasers and cameras for other applications, such as pavement condition surveys. Data are extracted using a data processing interface typically provided by the equipment manufacturer. Specialized expertise is beneficial in order to process and manage the LiDAR data effectively. Survey images can be used later to inventory a new asset (not identified prior to the survey) without resurveying. The process becomes increasingly cost-effectiveness as the number of applications for the data increases. Special Considerations Mobile LiDAR produces a significant amount of data so an agency might have to make special provisions to store the data. If used only for building a roadway asset inventory, many of the benefits to using this technology are not realized. Application Can be used to inventory all assets visible from the roadway. Data can be collected for multiple assets simultaneously at highway speeds. Value Added Assets can be located accurately to within a few inches. Vertical clearances and dimensions can be estimated within a few inches. Increases consistency by reducing human intervention. Point cloud can be reused for multiple purposes without another survey. Improves safety by reducing personnel in the field. Limitations LiDAR by itself does not capture objects in color, which may be used to classify some assets (such as signs). Assets that are not visible from the roadway surface cannot be inventoried using this method. Figure 4. Characteristics associated with mobile LiDAR. 11

Limitations: Only assets visible from the roadway surface can be inventoried using this method. The file size generated by the point clouds is large and may require agencies to make special provisions for data storage. Unless the technology is used for multiple applications that require the high degree of precision possible with mobile LiDAR, the full benefits of the technology are likely unrealized. Data generated from mobile LiDAR is gray scale, so additional capabilities must be added if color is used to differentiate some assets (such as signs). In summary, mobile LiDAR is a viable option for developing an asset inventory at traffic speeds to a high-degree of accuracy. Vertical elements and dimensions can be recorded without additional effort and the methodology can be used for applications beyond asset management. While some state DOTs have used mobile LiDAR successfully, its full benefits are realized when the data are used for a wide range of applications, including design applications. This method is limited to assets visible from the roadway, but can be combined with manual surveys to minimize this deficiency. SUMMARY Based on the information provided in this chapter, Table 1 summarizes the applicability of each of the three data collection methodologies for various roadway assets. Transportation agencies rarely collect data on a single asset at one time, so the most feasible methodology should consider all assets included in the inventory. ADDITIONAL READING MATERIAL Michigan Department of Transportation (MDOT). 2014. Monitoring Highway Assets with Remote Technology. Michigan Report Number RC 1607. Michigan Department of Transportation, Lansing, MI. National Cooperative Highway Research Program (NCHRP). 2013. Guidelines for the Use of Mobile LiDAR in Transportation Applications. NCHRP Report 748. National Cooperative Highway Research Program, Transportation Research Board, Washington, DC. Yen, K. S., B. Ravani, T. A. Lasky. 2011. LiDAR for Data Efficiency. WA-RD 778.1. Washington State Department of Transportation, Olympia, WA. FHWA. 2013. Iowa Department of Transportation's Tablet Asset Data Collection. FHWA. September, 2013: http://www.gis.fhwa.dot.gov/documents/newsletter_summer2013.asp Massachusetts Department of Transportation. 2013. Curb Ramp Inventory System. MassDOT. http://www.massdotinnovation.com/pdfs/session4mb-adaburbramp.pdf 12

Table 1. Applicability of each data collection methodology to inventory roadway assets. Roadway Assets to be Inventoried (F)easible/(P)referred Methodology Manual Photogrammetry LiDAR Signs Boards F P F Noise Barriers F F P Comments Photogrammetry captures color if needed to differentiate sign type Required dimensions lend themselves to use of mobile LiDAR Culverts P Not visible from the roadway Fences F P F Earth Retaining Structures F F P Accuracy offered by photogrammetric methods is sufficient Required dimensions lend themselves to use of mobile LiDAR Other Drainage Structures P Not visible from the roadway Guardrail F P F Concrete Barrier F P F Overhead Sign Structures F F P Pavement Markings F P F Accuracy offered by photogrammetric methods is sufficient Accuracy offered by photogrammetric methods is sufficient Vertical clearances best measured with LiDAR Photogrammetry captures color needed to differentiate markings ITS F P P Automated techniques save time Lighting F P F Accuracy offered by photogrammetric methods is sufficient Rest Areas P Not visible from the roadway Sidewalks F F P Tunnels F F P Required dimensions lend themselves to use of mobile LiDAR Required dimensions lend themselves to use of mobile LiDAR 13

CHAPTER 3 GUIDELINES There are many considerations that must be taken into account when selecting an approach for establishing a roadway asset inventory. The selection of the appropriate technology is only one of the considerations that has to be made. Prior to that decision, there are a number of important choices that have to be made regarding the assets that will be included in the inventory, the level of accuracy required, and the resources available to collect, process, and maintain the data. These decisions are influenced by many different factors, including the following: The importance or visibility of the asset For instance, it is more important to complete an inventory on a small number of highly-visible assets (such as tunnels) than on a small number of assets primarily installed for the convenience of the traveling public (such as recreational signs). The asset s role in reducing agency and user risk Transportation agencies face many risks in managing their assets from events such as natural disasters, financial uncertainties, and legislative changes. Building an inventory of high-risk assets may be an important strategy for helping to mitigate these risks. For example, some agencies have established inventories of areas susceptible to rock slides as a risk-mitigation strategy. An agency should also consider the potential risks to the users of the roadway if the assets are not maintained adequately. Certain types of assets, especially those in place to address safety concerns, are often the highest priority assets when establishing an asset inventory. The amount spent on maintaining the asset When prioritizing asset inventories, it may benefit the agency to evaluate the relative amount of money spent on maintaining an asset, or the total number of assets being maintained, as compared to the entire asset population. Assets that consume a large portion of the maintenance budget may be a higher priority for establishing an asset inventory than other assets. The relevance to the agency s strategic goals and objectives As agencies mature in their use of performance-based data for making investment decisions, it will become increasingly important that decision-makers have the information needed to support their strategic goals and objectives. Therefore, a higher priority may be established for building an inventory of assets that enable the agency to meet its performance targets. The availability of consistent protocols for collecting and storing data To help ensure that data are developed consistently throughout the agency, it is important that data collection protocols or guidelines are established for each asset that will be added to the agency s inventory. This step should be completed before steps are taken to build the asset inventory. The existence of regulations or other mandated requirements Transportation agencies have to adhere to certain regulations (e.g. retro-reflectivity standards, ADA compliance) that have been established by Federal, State, or other regulatory agencies. It is important to build and maintain the inventory of these assets to meet requirements, address safety concerns, and avoid negative consequences. The existence of legacy systems and processes Over the years, transportation agencies have implemented many software programs that are used to manage highway assets. The existence of these systems may influence the type of data to be collected, and the format that is used for storing the information. In today s organizations, there is an emphasis on 14

data integration of multiple data sources so the availability of geospatial references to link features is becoming increasingly important. Even after selecting a technology, the agency must decide how the data will be collected and whether an outside source will be used. Additionally, decisions regarding data processing and management will also need to be made, to help ensure that the data is used effectively by as many users as possible. The guidance provided in this chapter introduces the considerations that should be made at each step in the process. Examples of practice are provided throughout the chapter to help illustrate the points being made and supporting information is provided in the Appendices to help agencies better understand the type of information that is needed. The information is organized into the following steps, each of which is an important part of the decision process: Step 1: Getting Ready to Select a Methodology. Step 2: Selecting a Methodology. Step 3: Collecting the Data. Step 4: Processing and Managing the Data. The considerations that should be taken into account at each step of the process are illustrated graphically in Figure 5 and described throughout this chapter. Collectively, the information provided in this chapter guides the reader through the process of navigating these issues successfully while helping to establish a good understanding of the influence that each factor has on the usefulness of the data and the cost-effectiveness of the technology. STEP 1: GETTING READY TO SELECT A METHODOLOGY Even before discussing the options for selecting a methodology to be used for developing the inventory, an agency needs to identify its data needs, determine the characteristics of each asset that will be collected, and identify any financial constraints that may impact the methodology selected. These activities should be completed before any data collection efforts are initiated, regardless of the methodology that will be used. Select Assets to Include in the Inventory One of the first steps for an agency is to determine which roadway assets will be included in the inventory. This can be a challenging activity because there is often pressure from field personnel to establish inventories for all roadway assets. However, agencies just beginning to track asset inventory information may find it beneficial to start by adding a few high-profile assets and add to the inventory over time. There are a number of different approaches to take in selecting the assets to include in the inventory. According to a synthesis of practice, the most complete inventories in state DOTs (excluding pavements and bridges) include culverts, overhead sign structures, signs, signals, variable message boards, impact attenuators, pavement markings, guardrail end treatments, and rest areas (NCHRP 2015). Some states, such as Ohio and Nevada DOT, have conducted studies to assess the completeness of asset information, its contribution to agency decisions, and the risk associated with missing or incomplete data. The results of these studies have served as the basis for assigning a priority ranking to each asset. The asset inventories are then established for the 15

highest priority assets first. Assets related to an agency s safety goals can often be found in the top priority category. Figure 5. Guidelines for developing or updating a roadway asset inventory. 16

In some cases, the assets that are included will influence the manner in which the data is collected, as in the case of culverts that are not visible from the driving lanes. For the majority of the remaining assets, though, any of the three data collection methodologies is viable. While it is tempting to collect information on as many assets as possible, an agency should carefully consider its decision since the data must be maintained over time. If data collection and governance standards have not been developed for a particular asset, it is generally better to delay the data collection process until these steps have been completed. The number of assets included in the inventory will influence the amount of time required to collect the data if manual techniques are used. Since photogrammetry and mobile LiDAR both collect data at traffic speeds, data collection efforts are not influenced by the number of assets being collected. However, the data processing activities will likely be influenced by the number and type of assets included in the inventory. In general, the automated data collection methods can process data on a large number of assets efficiently. Determine Resource and Other Constraints Another set of factors that must be determined prior to data collection are the resource and/or contracting constraints that may influence the methodology selected. One example of a resource constraint includes the funding available for the data collection activity, both for the initial efforts and for future efforts to maintain the data. Another consideration is the availability of inhouse personnel to collect, process, and manage the data that is collected. Manual data collection techniques conducted by in-house staff represent a significant investment of manpower, so an agency considering this methodology needs to ensure that the data collection efforts will become a regular, on-going part of the workload for maintenance personnel. Photogrammetry and mobile LiDAR lessen the workload for in-house personnel, but introduce contracting requirements for obtaining a contractor or buying the equipment. Since both of these technologies are considered specialized services, agencies will have to verify that there are no contracting requirements to purchase locally for these services. The use of a contractor also introduces other issues that the agency will have to consider. For example, consistency in the data from year to year is extremely important since the results are used to influence investment decisions. This consistency can be impacted by changes in equipment, technology, and/or contractors. To minimize these impacts, some state DOTs have established multi-year contracts that include options for additional surveys in future years so that the same contractor, equipment, and technology can be used for several consecutive surveys. Identify Users The data collection activities associated with building a roadway asset inventory can often be conducted in conjunction with other existing activities, such as Maintenance Quality Assurance (MQA) inspections or pavement management surveys. Additionally, the results are often used by multiple divisions within an agency, including Maintenance, Operations, Safety, Traffic, Asset Management, Design, and Planning. Prior to initiating data collection efforts, it is important to identify the potential users of the data to a) identify their specific information needs, b) determine the data format that is needed to integrate into legacy software programs, c) establish the frequency for updating the information, and d) identify the resources available to support these efforts. The information obtained through this process will help to determine whether one methodology is more viable than another. In general, the more users available to 17

share the data and the costs associated with data collection and processing, the more costeffective automated techniques become. The uses for the data that will be collected has a significant impact on the level of detail that is needed from the survey. For instance, information that is used to make strategic decisions (e.g., agency goals) is typically less detailed than the information needed to identify projects and treatments. This concept is illustrated in Figure 6, which shows that as decisions move up in the organization, the level of detail and the quantity of data tend to decrease (Flintsch and Bryant 2009). As a result, an agency that simply wants to have an estimate of the number of signs needs much less detail than a maintenance supervisor who needs to know the type of sign and its location for scheduling maintenance activities. Figure 6: Relation between decision making levels and detail and amount of data required (Flintsch and Bryant 2009). Establish a Data Dictionary To help ensure that the data collected contains each of the asset attributes needed, it is recommended that 18 -------------------------------------------- Request for Proposal (RFP) Tip: Include a data dictionary in your RFP that defines the attributes you want identified for each asset. --------------------------------------------

each agency develop a data dictionary prior to selecting a methodology. This step is important to help ensure consistency in the data collection process and to verify that the data collection efforts will result in meaningful and complete information. A data dictionary describes, for example, the attributes that are to be collected, the level of detail required, and the level of accuracy that is expected. An excerpt from a data dictionary produced by the Tennessee Department of Transportation (TDOT 2014) is included as Appendix A. It illustrates the level of detail required to establish the inventory. If the work is being done by a contractor, the data dictionary is used by the contractor to prepare the project bid. STEP 2: SELECTING A METHODOLOGY Once data collection needs are understood and any constraints have been identified, an agency can begin the process of selecting the most appropriate methodology. In practice, there is no single methodology appropriate for all agencies. Rather, the selection of an appropriate data collection methodology is influenced by a number of different factors. These factors, and their influence on each of the methodologies, are discussed under this step of the four-step process. Evaluate Asset Visibility from the Road Of the three data collection methodologies included in this Guide, only manual techniques can be used to build an inventory for assets that are not visible from the traveling lanes of a road. However, the use of manual techniques for certain assets, such as drainage structures, does not prevent an agency from using another method of data collection for the other assets included in the asset inventory. For most other roadway assets, any of the three methods of data collection provide a viable option for establishing an asset inventory. According to a recent survey of practice, manual techniques are most commonly being used to build asset inventories (NCHRP 2015). However, as agencies streamline their data collection processes, they are exploring the opportunity to consolidate data collection efforts (as discussed later in this section). In general, if agency personnel will be establishing and maintaining the asset inventory and updating the inventory as work activities are conducted, manual data collection techniques will likely continue to be used heavily. However, the automated data collection techniques provide an opportunity for agency personnel to establish an inventory using data extraction programs provided by the vendor while sitting at a computer workstation. These options enable agency personnel to build the inventory, without requiring them to go out in the field. Additionally, if automated technology is already being used by the agency for other purposes (e.g., pavement management surveys), asset information can be collected while the surveys are being conducted and processed at a later point in time if that better serves the needs of the agency. Consider Accuracy Requirements The accuracy requirements for inventory data also have an influence on the methodology used for data collection. Manual techniques have the lowest level of accuracy, with location measurements considered to be accurate within a few feet. The accuracy of location measurements for photogrammetry is approximately 1 foot, but mobile LiDAR can be accurate to ±3 inches, if calibrated carefully. While the level of accuracy available with mobile LiDAR might not be important for most assets, there may be certain assets for which a more precise level 19 ------------------------------------- While the Utah DOT uses LiDAR for collecting most of its roadway asset data, its inventory of drainage assets and underground utilities was established by part-time interns using manual data collection techniques. -------------------------------------

of accuracy is important. For instance, it may be important to know the width of paved roadways to a high degree of accuracy if the information is used for developing project cost estimates. Guidance on establishing the level of accuracy required is provided as part of Step 3. Assess Agency Maturity Another consideration in selecting the appropriate data collection methodology is determining the maturity of the agency in terms of being able to fully utilize the data provided. Agency maturity takes into account several aspects of the data collection process. First, it is important to determine whether the data collected has applicability across the agency. For instance, mobile LiDAR can provide a large amount of detailed information; however, if that data is not fully used within the agency, photogrammetry may be a suitable alternative. A second aspect of maturity has to do with the agency s knowledge and understanding of each methodology. The more complex the use of technology, the more important it is to involve individuals with strong technical backgrounds in the selection process. For example, it may be important to include Information Technology in the selection process to ensure that the large files provided with the mobile LiDAR methodology can be managed by the agency. It may also be important to work with individuals to ensure compatibility with existing legacy systems that will use the data. For instance, involving individuals capable of working with the agency s Geographic Information System (GIS) is important for any methodology providing GPS coordinates. Finally, there may be specialized training needed by agency personnel to be able to work with the data obtained using one of the automated approaches since they involve the use of technology that is not familiar to all maintenance personnel. Consider Safety Requirements As discussed in the previous chapter, the use of mobile LiDAR and photogrammetry reduces the number of people who are collecting data in the field, which improves safety considerably. Measures can be taken to make manual techniques as safe as possible, but some agencies prohibit the use of manual survey techniques due to safety concerns. To significantly reduce the number of agency personnel in the field, either of the two automated methodologies should be used. Evaluate Resources ------------------------------------- Digital images from photogrammetry or mobile LiDAR could be used by Safety personnel to study the relationships between crash sites and the presence of crash barriers. ------------------------------------- Each of the three methods suggested for developing an asset inventory places different types of demands on agency resources. Manual techniques typically place the burden for data collection on individuals in the field, although contractors could be used. Photogrammetry reduces demand on individuals in the field, but there are mobilization costs associated with the data collection efforts, especially if data is collected by a vendor. Photogrammetry also requires resources to extract the data from the images, but this activity could be conducted by either agency personnel or a vendor. The resource requirements associated with data collection for mobile LiDAR are similar to photogrammetry, except that additional features are required on the van. Since most agencies do not own this equipment, mobile LiDAR is generally collected by a contractor. Resources are required to extract information from mobile LiDAR, but most agencies rely on the vendor to provide this service. 20

The relative cost, resource requirements, and utility of the data from each of the three methodologies is presented in Figure 7. As shown in the figure, mobile LiDAR has the highest data utility, since the information can be used for so many applications within a DOT. Manual techniques typically place the largest demand on agency resources, but this varies considerably based on the number of assets included in the inventory, the size and location of the agency, and the level of detail selected. The circles in Figure 7 represent the relative cost of each approach when taking into account the equipment and services being provided. Again, the actual costs to an agency should be carefully evaluated to determine a more realistic comparison. Figure 7: Relative comparison of resource requirements, data utility, and costs (not to scale). -------------------------------------------------------------------------------------------------------------- Automated Data Collection Costs The cost of collecting data using automated equipment varies considerably based on a number of factors, including the project location (e.g., impacting mobility costs), the size of the network, the number of different types of assets to collect, and the level of detail required. Several examples of automated data collection costs are documented in the literature, but the resulting costs are heavily influenced by the size and nature of the study (MDOT 2014 and Jalayer et al. 2014). These resources indicate that photogrammetry costs range from $72 to $88 per mile and mobile LiDAR costs range from $540 to $933 per mile. However, data collection vendors indicate that the cost of mobile LiDAR is dropping as the technology is developed and as it becomes more practical for transportation applications. In conversations with several state agencies, their anecdotal experience has shown that mobile LiDAR is about four times the cost of photogrammetry. -------------------------------------------------------------------------------------------------------------- Identify Other Data Collection Efforts Another consideration in selecting a data collection methodology involves an assessment of other data collection efforts that could be combined with the efforts to establish a roadway asset inventory. Most commonly, transportation agencies that are using photogrammetry for collecting pavement management data are adding cameras or mobile LiDAR to the equipment to broaden the applications associated with the data collection process. Combining ------------------------------------- Maryland DOT owns automated equipment for conducting pavement condition surveys that is also used for asset management purposes. ------------------------------------- 21

Manual Survey Photogrammetry Mobile LiDAR A Guide to Collecting, Processing, and Managing Roadway Asset Inventory Data multiple survey approaches can be very economical because it eliminates duplicative efforts and adds minimal additional cost to the original data collection efforts. Summary A summary of the key considerations in selecting a data collection methodology is provided in Figure 8. Fair degree of accuracy (± a few ft.) Labor intensive Safety issues with personnel in the field Quality control activities require additional personnel in field Best option for inventorying assets not visible from the road Does not require specialized technical expertise or equipment Most applicable when collecting a limited amount of data Good accuracy (± 1ft.) Not labor intensive Requires specialized equipment Operates at traffic speeds Can only be used to inventory assets visible from the road Easily used in conjunction with automated pavement condition surveys Data can be used by multiple Divisions within an agency Quality control activities can be done at a workstation Requires some technical expertise High degree of accuracy (± 3in.) Not labor intensive Requires specialized equipment Operates at traffic speeds Can only be used to inventory assets visible from the road Provides features for estimating asset dimensions Easily used in conjunction with automated pavement condition surveys Data can be used by multiple Divisions within an agency Quality control activities can be done at a workstation Provides greatest benefit when data are used by multiple Departments Requires specialized technical expertise Generates large data files that must be managed Figure 8. Factors in selecting a methodology for building a roadway asset inventory. STEP 3: COLLECTING THE DATA The next step in the process involves collecting the data using the methodology selected and periodically evaluating the process to determine whether a different method of data collection is warranted. This section of the Guide describes the activities typically conducted once the methodology for data collection has been determined. Secure Data Collection Equipment and/or Vendor 22

Once the data collection methodology has been selected, the agency must determine whether it has the personnel, equipment, and expertise required to collect the information using in-house resources. If an agency has elected to establish its inventory using an outside contractor, the next step in the process is to acquire the services of a data collection vendor through normal purchasing procedures. As a part of this process, agencies formulate and release Requests for Bids or Requests for Proposals. Since the content of the proposal request serves as the basis for the services that will be provided, an agency should spend considerable amounts of time developing the technical specifications that will be followed by the vendor. Several state transportation agencies have issued RFPs for data collection services that are available on their websites (UDOT 2011, TDOT 2014). A summary of the typical content included in a data collection RFPs is included in Appendix B. If the data will be collected using in-house personnel, the agency will have to acquire any equipment needed to build the asset inventory. If the agency has elected to purchase an automated data collection vehicle from a vendor, the agency may also need to obtain software licenses from the vendor so that the data can be processed by in-house staff or to allow viewing of the images at a workstation. Agencies that have used automated data collection for a number of years have found it beneficial to set up their data collection contracts for a multi-year period so that more than one data collection cycle is included. This approach helps to ensure the consistency of the data between cycles since the vendor will not have to repeat the learning curve that takes place each time a vendor works with a new agency. Some agencies include later data collection cycles as options so the agency has the opportunity to evaluate the vendor s performance before deciding whether to extend the contract. Agencies have also found it useful to reduce costs by limiting the number of software licenses purchased from the vendor. This requires the agencies to process their data in a single, central location and then distribute the data available to others in the agency in a format that is not tied to proprietary software. Alternatively, statewide licenses can be obtained so that any potential user of the data has the tools necessary to view the data at a workstation. However, there are significant training requirements associated with this approach that have to be planned for and that may be difficult to maintain over time. Develop Data Collection Protocol It is important for agencies to establish a well-defined data collection protocol to be used to help ensure consistency in the results. The protocols can be documented in a data collection manual that can be used by field personnel or data collection vendors to describe details about the data collection process and to outline the steps that will be taken to ensure the quality of the data. Guidance on developing a quality management plan for automated data collection activities associated with pavement management are available in the literature (Pierce et al. 2010). Although the documentation outlines considerations for monitoring pavement quality, many of the same considerations should be incorporated into a data collection protocol for establishing a roadway asset inventory. At a minimum, the quality management plan should include the following (Pierce et al. 2010): 23 ---------------------------------------- RFP Tip: Agencies using automated data collection vendors have found it helpful to establish a contract period that covers at least two data collection cycles to help ensure consistency. For instance, an agency may establish a contract for one data collection cycle with an option to renew the contract for another cycle if the agency is satisfied with the vendor s performance. ----------------------------------------

The deliverables that will be provided, the protocols that will be used for collecting the data, and the required resolution, accuracy, and repeatability to determine the quality of the data. The quality control activities that will be conducted and how frequently they will be performed. In general, if data collection is being performed by an outside vendor, the vendor is responsible for monitoring the quality of its processes, but the agency should verify that the vendor has a plan in place and is following the plan. The acceptance testing that will be performed to determine whether quality criteria have been met and the corrective actions that will be taken if deliverables do not meet the criteria. An agency is responsible for acceptance testing, regardless of whether the data are collected by agency personnel or an outside vendor. Simple acceptance testing should verify completeness of the data and the reasonableness of the values provided. More complex acceptance testing includes tests to verify a portion of the data provided, either in the field or at a workstation. Acceptance testing on approximately 5 percent of the network is typical (Pierce et al. 2010). Roles and responsibilities for each participant in the data collection process. Plans for documenting the quality management activities. A signature page verifying that each of the parties is familiar with the quality management processes and understands his/her roles and responsibilities. The data collection protocols provided to the field crews or to a vendor should provide enough specificity to help ensure that the data collection process proceeds as planned. For roadway asset inventory data, the level of detail provided in the data dictionary included as Appendix A is a good start. Additional information to guide the data collection process, such as maps showing the locations of all roads and ramps included in the survey and data formatting requirements, may also prove to be important to the protocol. A trouble-shooting guide may also be a valuable resource to personnel in the field. Conduct Personnel Training and Equipment Calibration Depending upon the choice of methodology and whether the agency decides to perform the data collection in-house or by contract, the type and amount of training will vary. Training of data collection personnel is important for quality control to help ensure consistency between raters and between survey years. For manual surveys, training is a common activity used to ensure the quality of data, but the availability of data collection manuals is also heavily relied on (NCHRP 2015). For agencies using automated equipment, it is important that the data collection crews know how to calibrate, operate, and troubleshoot the equipment. Some agencies regularly certify that data collection personnel have the necessary skills and knowledge to collect the data accurately (NCHRP 2015). If automated equipment is used to collect the data, it is important that the equipment is calibrated prior to the start of the surveys and periodically during the data collection process. Calibration sites may be established by the agency to verify that the data collection process is working as planned immediately prior to the start of surveys. Agencies using calibration sites do not allow the formal data collection processes to begin until the equipment has performed acceptably. During the data collection process, blind sites may be established to further verify the data being 24

collected. A blind test site is a location that is known to the agency, but not to the individuals responsible for collecting the data. Conduct Quality Control and Acceptance Testing During the data collection process, it is necessary to periodically monitor the data being collected in accordance with the quality control processes established prior to the start of data collection. This type of testing helps reduce the possibility of errors by identifying malfunctioning equipment, anomalies in the data sets, or other types of faulty or missing data. The parties responsible for collecting the data are responsible for performing quality control testing, so this is a vendor s responsibility if an outside contractor is used. If a manual process is used, the agency is responsible for conducting any quality control checks that might be needed. Acceptance testing is the responsibility of the data owner (the agency). It involves performing checks on the data provided by the data collection team to verify that it meets the established standards for quality. At its simplest level, acceptance testing is used to verify the completeness of the data and the reasonableness of the values provided (e.g., they fit within established ranges), but it can also involve manual checks of a small percentage of the data to verify the accuracy of the data before accepting it into a database. Acceptance testing on 5 percent of the network is used by some agencies (NCHRP 2015). STEP 4: PROCESSING AND MANAGING THE DATA The final step in the process involves the activities associated with extracting the information from the data and presenting it in a format that can be used to support agency decisions. Develop In-House Technical Expertise When adopting any type of new technology, it is important that the users of the data are trained sufficiently so they understand its capabilities and limitations. With manual data collection processes, the technical expertise likely resides in the agency already. However, the automated technologies described in this Guide likely involve new equipment and data processing techniques that may not be familiar to agency personnel. Even if a contractor is used to process the data, it is important for agency personnel to be trained sufficiently to be able to conduct quality acceptance testing and to use the data fully. If the agency is processing the data in-house, additional training may be needed to master the skills involved with operating the data extraction software. Formulate Data Processing Procedures The data collection process for a large transportation network generates a large amount of data. To assist with the processing of the data, data collection vendors have developed software that allows them to process the information efficiently. The software is frequently licensed to transportation agencies to facilitate their use of the data at a computerized workstation. In some cases, the transportation agency may elect to process the data collected using photogrammetry with in-house personnel viewing the images on a workstation within the software obtained from the vendor. Simple point-and-click routines perform the necessary identification and storage of the data, so the process can be done relatively efficiently. However, the more assets that are included in the inventory, the longer the data processing can take. Other agencies contract with the data collection vendor to perform the data extraction and the vendor may elect to use either automated or manual processes, depending on the method used to collect the data. If the vendor 25

processes the data, it is good practice for the agency to perform acceptance testing on a small sample of the data submitted to ensure that the data requirements have been met. If processing is to be done manually at a workstation with in-house personnel, it may be a good idea to incorporate the following suggestions into the data processing procedures: To help ensure that quality doesn t suffer, limit the amount of time in front of the workstation to 4 hours a day, with no consecutive block more than 2 hours in length. Consider establishing processes that prevent in-house personnel from processing data in an area where they are responsible for the maintenance of those assets. This helps to ensure that the surveys are conducted by an independent party. Identify an independent rater to check the accuracy of the ratings on randomly-selected samples of the network. Provide Access to Data Once the data is collected and all necessary information has been extracted, it is important that the information is made available to other potential users in an easily accessible format. The use of Excel and Access datasheets is common for data collected manually. The data sets associated with automated data collection are often very large and may need to be condensed to a more manageable size for others to use the data. If the agency acquires a statewide license for the workstation viewing software, most users could access the images at a workstation. However, this may require training on an on-going basis, so the agency should make provisions for this. A statewide viewing license is not the only way to make data available to users since images can be linked to other programs that are more familiar to agency personnel. In some cases, agencies have required a data collection vendor to deliver the processed data with a graphical interface that will enable personnel from throughout the agency to view the information. An example of the type of interface provided to the New Mexico DOT is provided in Figure 9. The advantage to this type of graphical interface is that it makes it easy for a user to find the particular information of interest through a map or list interface. Studies have shown that minimizing the number of hurdles that have to be overcome to access the data leads to increased use of the data and greater value to the organization (Hensing and Rowshan 2005). 26

Figure 9. Screenshot of New Mexico RFI spatial map with assets identified (Hensing and Rowshan 2005). Address Organizational Issues For most agencies first establishing their roadway asset inventory, there are important organizational issues that need to be addressed in order to utilize the information fully and to help ensure that the information remains current and relevant to the agency s business processes. For agencies electing to use manual techniques for collecting data, there are likely to be very few organizational issues that will need to be addressed for the activity to be successful. On the other extreme, agencies electing to use mobile LiDAR should develop business processes that promote the use of the data outside of Maintenance or Asset Management to fully realize the potential benefits. This often means reaching across traditional organizational stovepipes to ensure that the information collected through this process can be used fully. This may require the development of processes to ensure that data conforms to the needs of existing legacy systems, including maintenance management systems, construction management systems, and safety management systems. Implement Data Governance Standards As the agency s data collection activities advance, and especially as re-inspections are conducted, it is important for the agency to monitor its efforts in accordance with its data governance standards. These standards, which identify the attributes to be collected for each asset type, also specify the data owner as well as all users of the data. This information is very important in successfully managing the inventory data. It helps to ensure that changes to data attributes and formats that could impact legacy software programs are not made inadvertently. Develop Plans for Inventory Updates Ideally, the content of the roadway asset inventory is updated regularly as maintenance personnel, contractors, and other field crews perform work in the field. Periodically, it may be beneficial to conduct an update to the asset 27 ---------------------------------------------------- Sample state data collection approaches and inventory cycles * North Carolina DOT Photogrammetry Updated every 3 to5 years * Maryland SHA Photogrammetry Updated every 5 years * Tennessee and Utah DOTs LiDAR Updated every 2 years ----------------------------------------------------

inventory to replace outdated information and to verify the accuracy of the inventory data. In some instances, rather than establish business processes to update the inventory data, old inventories are purged and replaced by newer data each time a survey is completed. This approach is only feasible if the inventory data is not stored by asset in a maintenance management database. Instead, agencies using this approach generally work from a count of assets rather than manage each asset individually. Agencies should take steps to ensure the quality and consistency of data from one inventory cycle to another. Some agencies have used multi-year contracts that cover multiple data collection cycles as a strategy to improve data consistency. Other Considerations In addition to the considerations already mentioned, there are several additional issues that should be noted, as discussed below. Before a second cycle of data is collected, the agency must determine whether the previous inventory will be replaced or whether the new information will be used to update the previous data. The latter approach is used when an agency maintains a database for each asset in the asset inventory. A technique referred to as ghosting has been used by some agencies to compare two data sets from different inspection cycles as a process for updating the inventory. Ghosting allows agencies to see changes in the data files from one data collection cycle to another. If the data files are merely replaced, individual changes in the inventory cannot be identified, although total changes in the number of assets can be determined. When data is collected using either photogrammetry or mobile LiDAR, the images or the point-cloud can be used at any point in the future to extract inventory information, or to add an asset not extracted initially. This feature is very beneficial because it allows the agency to expand its inventory without requiring additional field work or equipment mobilization. Pavement condition surveys continue to rely primarily on digital images from downwardfacing cameras associated with photogrammetry equipment. This same equipment can be used to develop an asset inventory with only minimal adjustments to include additional cameras. Mobile LiDAR can also be added to a van equipped with photogrammetry, if desired. Mobile LiDAR on its own has not been used for pavement condition surveys due to limitations in the resolution of its point-cloud. The point-cloud data generated by mobile LiDAR is so significant in size that agencies must plan how to manage the data sets. One state reported that its statewide inventory files reached seventeen terabytes of storage and cost ----------------------------------------- It cost one state transportation agency $90k/month to store the point-cloud data from one cycle (17 TB) for a network of about 28,000 lane miles. ----------------------------------------- approximately $90,000 a month for storage on state servers. The contract covered approximately 28,000 lane miles. The mobile LiDAR point-cloud is a digital elevation model (DEM) that is very beneficial for transportation design and planning applications. To fully realize the benefits associated with mobile LiDAR, these types of applications should be explored. 28

EXAMPLE Otherwise, unless a high-degree of accuracy or dimensions are needed, photogrammetry may be a suitable substitution. The decision regarding the use of manual or automated approaches to establish a roadway asset inventory requires the consideration of many factors. These factors have been organized into a series of four steps that include decisions that have to be made regardless of the technology, those that will influence the selection of the most appropriate technique, and the remaining considerations that have to be accounted for while collecting, processing, and managing the data. To illustrate how these factors contribute to the selection of a technology, an example is provided. The example is completely hypothetical and is intended only to illustrate the use of the four steps to address all of the considerations involved in establishing and maintaining a roadway asset inventory. The Scenario A fictional Department of Transportation, known as XDOT, has seen the number of fatalities due to crashes rise in the past 5 years. The XDOT Safety Division is concerned about the increase in fatalities and the Legal Department has also warned executive leadership about possible lawsuits if the trend continues. The agency has determined that having information about its safety assets would help XDOT address safety-related deficiencies. In particular, XDOT is interested in correlating crash locations with the location of guardrails and message boards. Since there is no inventory information on these assets, the Safety Division has identified the establishment of a guardrail and message board inventory as a high priority. The Process Step 1: Getting Ready to Select a Methodology In addition to the needs of the Safety Division to georeference the location of guardrails and message boards, the Maintenance Division expressed interest in collecting information about the type of guardrails and message boards in place so they can establish a maintenance schedule for these assets. The Department also discovered that the Pavement Management Division has been using equipment outfitted with photogrammetry to collect information on pavement conditions each year. Once the users had been identified, representatives from Safety, Maintenance, and Pavement Management met to establish the inventory characteristics that would be collected on the guardrails and message boards. The data dictionary that was referenced in Appendix A of the Guide was very useful in establishing the level of detail that would be needed. The information was incorporated into a data dictionary and Maintenance was assigned responsibility for storing the data since a Maintenance Management System had recently been implemented. Step 2: Selecting a Methodology Each of the three methodologies presented in the Guide was considered to be a viable option for collecting the inventory information on guardrail and message boards since both assets are viewable from the travel lanes. The users identified the location of the assets, the length of the guardrail, and the type of guardrail or message board as the most important information to be obtained from the process. After much discussion, the group decided that it was not necessary to determine the height of the guardrail as part of this process. Therefore, the group decided that a location accuracy of 1 to 2 feet was acceptable. 29

Since the inventory was being established to reduce safety hazards, XDOT was hesitant to require agency personnel to collect the inventory data in the field. The fact that a viable data collection method was being used for pavement management purposes convinced the agency representatives to select photogrammetry as the preferred methodology. To help ensure that the data was processed as quickly as possible, XDOT elected to use the contractor to extract the inventory information from the first run; however, the Department asked the vendor to price data extraction as an optional cost in case agency personnel were interested in extracting data in a future survey. Step 3: Collecting the Data Since photogrammetry was already being used for pavement management purposes, the agency elected to advertise for a new data collection contract that included ROW and side-oriented cameras to collect the guardrail and message board information. Individuals from Maintenance and Safety worked with the pavement management team to learn how to use the workstation for viewing asset data. In addition, a quality plan was developed to help ensure that the data was collected in accordance with the standards outlined in the data dictionary. The Request for Proposals was advertised and a vendor was selected and a contract signed. The cost of the additional data collection and processing was nominal and XDOT was pleased that the data could be obtained so cost-effectively. The data collection process began with equipment certification and, upon approval, the vendor was authorized to begin the data collection process. Data would be submitted in batches, closely following the boundaries of each District within the State. Step 4: Processing and Managing the Data To minimize the amount of time required for processing the inventory data, the guardrail and message board information was extracted by the vendor using proprietary software they had developed. The contract with the vendor provided two licenses for workstations, with one being placed in the Safety Division and the other placed in Maintenance. Maintenance personnel were trained in data extraction by the vendor and two central office maintenance personnel were assigned responsibility for using the workstations to check the results of at least 5 percent of the data provided by the vendor. The vendor also provided the data in a format that could be made easily accessible to other personnel within the Department without using a workstation. The georeferenced data were linked to the agency s GIS map so it could be compared with crash locations. Since a multi-year contract was established with the vendor, XDOT is assured that the inventory will be updated on a 2-year cycle. The Maintenance Division expects to investigate the feasibility of extracting additional information from the images as they become more familiar with the technology. They have decided that signs and light structures will be their next highest priority. Since the images and workstations are available to XDOT, the identification of signs and light structures can be done at any time following the development of the data dictionary. ACCELERATING THE LEARNING CURVE During the process of developing the Guide, a number of state highway agencies shared their experiences with building a roadway asset inventory using either manual or automated processes. The information has been compiled into two categories: a) challenges and possible remedies, and b) benefits realized. 30

Challenges and Possible Remedies For some agencies, drafting the RFP was a challenge because of the lack of technical expertise or prior experience with the technology. Opportunities for sharing RFPs and discussing both positive and negative experiences with peers has helped overcome this hurdle. Copies of the Tennessee and Utah DOT RFPs can be accessed online using the following links: http://tn.gov/generalserv/cpo/sourcing_sub/documents/40100-40914.pdf https://www.udot.utah.gov/public/ucon/uconowner.gf?n=11823602292354098 Acquiring, setting up, and learning new software has been a challenge for some agencies, but including training for in-house personnel as part of the deliverables has helped overcome this issue. There is little guidance available regarding acceptable levels of accuracy and precision for automated data collection efforts. The development of guidance in this area, as well as training on quality management activities would be helpful. In cases where agencies have transitioned from one data collection methodology to another, they have had problems with populating their legacy software programs with the new data unless specific steps are taken to address these issues. One of the challenges concerns differences in the level of accuracy related to size and geospatial relationships so that legacy systems can use the new data. In some agencies, it has been a challenge to keep the inventory up to date because of constrained resources. In some agencies that are building and maintaining inventories manually, the work may not get done because of other demands on time. Some agencies have elected to update their asset inventory on a 2- or 3-year cycle and the old data is purged when the new data is received. This approach works for agencies that do not track individual assets, but it would not work if an agency has a database that tracks historical data for each asset individually. According to information provided in the literature, the cost of LiDAR is significantly higher than photogrammetry options. However, as the technology develops and becomes more common in transportation agencies, the costs are expected to drop. In some cases, the full advantages associated with the use of automated technology have not been realized because the data has not be leveraged across the agency. In the future, it may be hard to justify data collection costs unless the data can be shown to address multiple needs within the agency. The more exposure the data has within the agency, the more valuable it becomes. The data sets produced by automated data collection techniques, especially LiDAR, are extremely large. In some transportation agencies, data storage is a centralized government function and DOTs are charged for the amount of storage used. To avoid excessive data storage costs, some data collection vendors offer to store data on their servers and provide client access through the cloud or through network/vpn links. Benefits Realized The use of automated technology features in this Guide has resulted in several benefits to transportation agencies, including the following. Improved safety by removing personnel from the field. 31

Improved organizational efficiency since multiple data needs can be addressed in one data collection effort and data can be checked at a workstation rather than sending individuals to the field. Coordinated data collection efforts reduce duplication of efforts. Leveraged data helps reduce the cost of data collection. Reduced agency risks due to improved access to asset information. Improved network conditions since agency priorities can be better addressed. Enhanced communication to better convey funding needs and/or enhance accountability. ADDITIONAL READING MATERIAL National Cooperative Highway Research Program (NCHRP). 2003. Quality and Accuracy of Positional Data in Transportation. NCHRP Report 748. National Cooperative Highway Research Program. Transportation Research Board, Washington, DC. National Cooperative Highway Research Program (NCHRP). 2007. Managing Selected Transportation Assets Signals, Lighting, Signs, Pavement Markings, Culverts, and Sidewalks. NCHRP Synthesis 371. National Cooperative Highway Research Program. Transportation Research Board, Washington, DC. Pierce, L. M., G. McGovern., and K. A. Zimmerman. 2010. Practical Guide for Quality Management of Pavement Condition Data Collection. U.S. Department of Transportation. Federal Highway Administration, Washington, DC. 32

CHAPTER 4 FUTURE DIRECTIONS Over the past 10 years, there have been tremendous advancements in the technology available to support efforts to establish and update roadway asset inventories. As a result of these advancements, the automated data collection techniques described earlier in this Guide are being used more frequently to support infrastructure asset management and maintenance management activities. Technology is expected to continue to advance, which will lead to further changes in the methodologies used to update asset inventories in the future. This chapter introduces some of the anticipated changes that may impact future efforts to build asset inventories. In addition, it documents the importance of reviewing data collection efforts regularly to ensure that the right information is being collected in the most efficient and effective method possible. FUTURE MODIFICATIONS TO THE DATA COLLECTION PROCESS Because of resource limitations, most agencies prioritize their data collection efforts to help ensure that the most important information is available to support existing business processes. As a result, few agencies are able to collect data on all of the assets that they are responsible for operating and maintaining. However, as transportation agencies become more comfortable with their data collection efforts, or as the number of assets managed using performance-based decision processes increase, it is likely that some agencies will elect to add to their roadway asset inventory at various points in time. Other changes to the data collection processes may be caused by further resource constraints that force an agency to revisit its existing methods of collecting asset inventory data and evaluate whether alternate approaches would be beneficial. Either of these situations illustrates the importance of periodically revisiting the steps outlined in the Guide to evaluate whether a different methodology might be warranted. For instance, an agency that is using photogrammetry to develop an inventory for guardrails, signs, and highway lighting may elect to add additional assets to its roadway inventory at some point in the future. The addition of some assets, such as pavement markings and protective barriers, could easily be added to the list of assets inventoried using the existing survey technique. And, depending on the amount of time that has passed since the data was last collected, it is possible that previously-collected survey data could be used to establish the initial asset inventory. However, if the additional assets to be collected include assets with accurate dimensional measurement that would benefit from the use of mobile LiDAR, the agency may want to change its data collection methodology for these assets. The use of mobile LiDAR could also be promoted by other business processes that want to build on the availability of the data collected as part of the roadway inventory. For example, if an agency intends to collect data on bridge clearances, mobile LiDAR provides an effective method of obtaining this information. It is suggested that agencies establish regular intervals for evaluating their data collection practices to determine whether alternate methods should be considered. This interval should be established based on the frequency with which agency practices change, the rate at which technology has evolved, and industry experience with each alternate approach. 33

EMERGING TECHNOLOGIES As technology continues to advance, a number of technologies that are currently either in the development phase or undergoing feasibility testing may become viable additions to the methodologies considered in the guide. Some of the more promising technologies available are presented in the remainder of this section. 360-Degree Camera One emerging technology is the development of a camera that consists of six lenses, positioned complementary to one another, providing a 360-degree horizontal perspective of an area. The camera uses mathematical algorithms to stitch together the various images to create the 360- degree view of the roadway. One of the six cameras is positioned vertically rather than horizontally to create a spherical view. These images, as in photogrammetry, are linked with GPS coordinates to identify field locations of the extracted data. This technology may provide additional benefits to traditional photogrammetry techniques in situations that would benefit from a 360-degree perspective, such as intersections and interchanges. Flash LiDAR In contrast to mobile LiDAR in which every single point is illuminated individually with a laser, flash LiDAR illuminates a whole scene at once. As a result, each pixel provides an indication of the amount of time that passed for the camera s laser flash pulse to hit the targeted asset and bounce back to the camera s focal plane. The time measurements are resolved using the speed of light, resulting in a 3-D image from the depth measurements for each point. Flash LiDAR is currently being tested for applications in the military and automobile industry due to its ability to provide real-time information. Flash LiDAR is also referred to as time-of-flight (TOF) cameras. Airborne LiDAR Aerial or airborne LiDAR has been around for several years but its use has been limited due to Federal Aviation Agency (FAA) flight restrictions that were imposed to avoid any conflicts with air traffic. Airborne LiDAR captures data on a scale that lends itself more to design and planning activities rather than building roadway asset inventories. For example, scans from airborne LiDAR have been used to create 3-D models of complex objects, such as piping networks, roadways, archeological sites, buildings, and bridges. Airborne LiDAR has also been used on large, civil engineering projects to assist with grading, utilities, drainage analysis, erosion control, and roadway design. It has also been used by the military and in the archaeological and agricultural fields. There are several advantages to airborne LiDAR that make the technology appealing to the transportation community. For example, objects can be measured remotely without interfering with traffic. In addition, the equipment can be operated under a variety of weather conditions and its sensors are not affected by low sun angles. Airborne LiDAR can even be used at night. In the past year, the FAA has granted approval to four companies to fly commercial drones to conduct aerial surveys, monitor construction sites, and inspect oil flare stacks (USA Today 2014). The results of the trials are expected to influence the future use of this technology. 34

Driverless Cars Automobile manufacturers have increasingly shown interest in the concept of driverless cars and the everyday use of this technology could soon become a reality. If that were to be the case, transportation agencies may have to shift their data collection priorities since driverless cars could require different roadway features to operate effectively. For example, striping could become increasingly important to keep the cars in the driving lane. Transportation agencies will face a major paradigm shift in terms of data collection and asset performance as this new technology becomes more common. ADVANCEMENTS IN DATA PROCESSING TECHNIQUES In addition to advancements in the methodologies being used to obtain asset inventory information, there are enhancements being developed for processing photogrammetry and mobile LiDAR data that are expected to benefit the transportation industry. For instance, a number of researchers are exploring the use of automated data extraction processes to obtain information on signs and other features from the images collected in the field. One particular application that shows promise is an automated matching and change detection technique that compares different data sets to identify changes (Habib and Al-Ruzouq 2012). This technique could be used to compare data sets from data collection efforts in different years so that changes to the asset inventory can be tracked with time. Initial efforts to use this technique have had limited success, but continued enhancements may make it viable in the future. Additionally, some agencies are exploring techniques for extracting asset features from LiDAR using ArcGIS software. The initial applications do not provide the same level of quality provided by existing extraction software, but continued efforts may improve the viability of this technology in the future. SUMMARY The establishment of an asset inventory is an important step in supporting an agency s asset management practices. Different methodologies are used to collect inventory information, ranging from manual techniques that involve the use of personnel who are directly involved in the measurement, to automated techniques that use noncontact sensor information and cameras to collect the data. Each of the different methodologies has advantages and disadvantages associated with it. Because of to the increasing importance of establishing asset inventories and the changes in technology that have been taking place over the last several years, there was a need for developing this Guide to provide a practical basis evaluating the options associated with collecting, processing, and managing roadway asset inventory data. As outlined in this Guide, there are a large number of factors that must be considered when determining the most appropriate methodology for establishing or updating an asset inventory. Further, the changing needs of the agency as well as the continued advancement in technology support the necessity for a regular assessment of the agency s data collection needs and the most appropriate method for obtaining that information. 35

ADDITONAL READING MATERIAL Cheok, G. S., M. Franaszek, I. Katz, A. M. Lytle, K. S. Saidi, N. A. Scott. 2010. Assessing Technology Gaps for the Federal Highway Administration Digital Highway Measurement Program. Internal Report 7685. National Institute of Standards and Technology, Gaithersburg, MD. Wang, K. C. P., Z. Hou, W. Gong. 2010. Automated Road Sign Inventory System Based on Stereo Vision and Tracking. Journal of Computer-Aided Civil and Infrastructure Engineering. Vol. 25, No. 6. John Wiley and Sons. pp. 468-477. Habib, A. F. and R. I. Al-Ruzouq. 2012. Linear Features for Automatic Registration and Reliable Change Detection of Multi-Source Imagery. Journal of Spatial Science. Vol. 57, No. 1. Taylor and Francis. 36

REFERENCES Flintsch, G. W. and J. W. Bryant. 2009. Asset Management Data Collection for Supporting Decision Processes. U.S. Department of Transportation, Federal Highway Administration, Washington, DC. Habib, A. F. and R. I. Al-Ruzouq. 2012. Linear Features for Automatic Registration and Reliable Change Detection of Multi-Source Imagery. Journal of Spatial Science. Vol. 57, No. 1. Taylor and Francis. Hensing, D. J. and S. Rowshan. 2005. Roadway Safety Hardware Asset Management Systems Case Studies. U.S. Department of Transportation, Federal Highway Administration, Washington, DC. Jalayer, M., H. Zhou, J. Gong, S. F. Hu, and M. Grinter. 2014. A Comprehensive Assessment of Highway Inventory Data Collection Methods. Journal of the Transportation Research Forum. Vol. 53, No. 2. North Dakota State University, Fargo, ND. pp. 73-92. Michigan Department of Transportation (MDOT). 2014. Monitoring Highway Assets with Remote Technology. Michigan Report Number RC 1607. Michigan Department of Transportation, Lansing, MI. National Cooperative Highway Research Program (NCHRP). 2007. Use of Mobile LiDAR in Transportation Applications. NCHRP Report 748. National Cooperative Highway Research Program, Transportation Research Board, National Research Council, Washington, DC. National Cooperative Highway Research Program (NCHRP). 2015. Maintenance Quality Assurance Field Inspection Practices. NCHRP Synthesis 470. National Cooperative Highway Research Program, Transportation Research Board, Washington, DC. Pierce, L. M., G. McGovern., and K. A. Zimmerman. 2010. Practical Guide for Quality Management of Pavement Condition Data Collection. U.S. Department of Transportation, Federal Highway Administration, Washington, DC. Tennessee Department of Transportation (TDOT). 2014. Request for Proposals for Statewide Roadway Asset Data Collection. Tennessee Department of Transportation, Nashville, TN. Accessed online from: http://tn.gov/generalserv/cpo/sourcing_sub/documents/40100-40914.pdf USA Today. 2014. FAA Lets Four Companies Fly Commercial Drones. USA Today. Accessed December 10, 2014: http://www.usatoday.com/story/money/business/2014/12/10/faadrones-trimble-vdos-clayco-woolpert-amazon/20187761/ Utah Department of Transportation (UDOT). 2011. Roadway Imaging/Inventory Program Bid Document. Utah Department of Transportation, Salt Lake City, UT. Accessed online from: https://www.udot.utah.gov/public/ucon/uconowner.gf?n=11823602292354098 37

APPENDIX A SAMPLE DATA DICTIONARY Attenuators Energy absorbing barriers which provide protection from vehicles striking rigid bodies such as bridge columns and barrier walls. LRS to reference GPS to reference Feature Type 04 Feature Char Choose from the following types Log mile location of front nose of attenuator. GPS location of front nose of attenuator. 00758 - GREAT 00759 - TRACC 00760 - Quadguard 00761 - Hex-foam Sandwich 00762 - React 01330 - TAU-II 01331 - SCI 01332 - HEART 01333 QUEST Feature Location Choose from the following locations 1-Left 2-Right 4-Median Right 6-Median Left 7-Centerline Height Width Report tallest height to nearest 0.1 feet. Measure vertically from the ground to the top of the attenuator Report widest point to nearest 0.1 feet. Length Report length to nearest 0.1 feet. Measure linearly along the centerline of the attenuator from the front nose to the point where the attenuator connects to A-1

the rigid body that it is protecting. Notes Only permanent installations shall be inventoried. Attenuators used for construction shall be excluded. Attenuators located along the center of the roadway or at the end of median barrier walls shall be coded as Feature Location = 7-Centerline. If an attenuator is found that does not match any of the examples provided, contact the State Project Manager. Flat-Sheet Signs A roadway sign which is fabricated using thin aluminum sheeting and a reflective sheeting to display directions and instructions to drivers. Flatsheet Signs are normally less than five feet in either width or height and do not contain reinforcing ribs on the back side. LRS to reference Log mile location of the sign. GPS to reference Feature Type 06 For ground-mounted signs, GPS will reference edge of sign closest to roadway (signs in right shoulder reference bottom left edge of sign, signs in left shoulder reference bottom right edge of sign). For overhead signs, GPS will reference center of bottom edge of sign. Feature Char Feature Location Choose from the following types The 5-digit code that corresponds to a specific MUTCD code will be entered (to be provided by the State) 1-Left 2-Right 3-Overhead Right 4-Median Right 5-Overhead Left A-2

6-Median Left 7-Centerline Feature Condition Choose from the following conditions 1-Poor: Sign may be damaged or non-reflective to the point that it cannot be clearly read by traffic. It may also be out of plumb enough that it is not legible. 2-Fair: Sign is clearly visible, mostly reflective, may have minor damage that does not interfere with the intended message of the sign, may be out of plumb, but still readable by traffic. 3-Good: Sign is clearly visible, reflective, free of damage, and plumb. Sign Orientation Choose from the following Sign Mount Type Choose from the following 1-North 2-South 3-East 4-West 5-Northeast 6-Northwest 7-Southeast 8-Southwest 01 Grnd Single Post-U shape 02-Grnd Single Post-square tube 03-Grnd Double Post-U-shape 04-Grnd Double Post-square tube 05-Grnd Double Post-W-beam 06-Grnd Triple Post-U-shape 07-Grnd Triple Post-square tube 08-Grnd Triple Post-W-beam 09-Bridge Mounted 10-Cantilever Overhead 11-Truss Bridge Overhead (This will include normal truss bridges intended solely for signs) Comments Enter the sign legend (Anything not denoted by the MUTCD code such as speed limit value for speed limit signs or other text like town names). Height Report height of sign to nearest 0.1 feet. Use the standard size that matches closest to the measurement; otherwise record to the closest inch. All measurements will be converted to feet before delivery. A-3

Width Notes Report width of sign to nearest 0.1 feet. Use the standard size that matches closest to the measurement; otherwise record to the closest inch. All measurements will be converted to feet before delivery. Legend that is written in the comments will follow rules put forth in Appendix A to ATTACHMENT E. Digital signboards will not be extracted. Only permanent sign installations will be collected, construction signs will be excluded. A-4

APPENDIX B TYPICAL CONTENT IN A DATA COLLECTION RFP The guide provides links to Requests for Proposals that were issued by the Tennessee and Utah DOTs for building their roadway asset inventories using automated techniques. Since publishing the RFPs in full is prohibitive, a summary of the key technical portions of a data collection RFP is provided here. In practice, the RFPs that were used in building this summary contained requirements for collecting both roadway asset inventory information as well as pavement condition data for pavement management surveys. Because of this, the RFPs may contain more detail than is required if the contract had been issued only for the roadway asset inventory. However, agencies realize the greatest benefits from the use of automated technology when multiple agency needs are addressed, so this was not considered to be a major issue. Agencies are encouraged to carefully consider their data collection goals and objectives, their data needs, and available resources when using this information to develop an RFP. The introductory sections of an RFP typically include the following. Reason for issuing the RFP and the goals it is intended to accomplish. Background information on the current state of the agency and its asset inventory. Information about the procurement process and pre-bid meeting. Intended length of the contract and the price guarantee period. Regulations on partnering, joint ventures, use of sub-contractors and so on. Terms and conditions relating to insurance, auditing, contract issuance and other agency policies. The technical specifications outlined in the body of the RFP typically includes the following types of information. General information on mileage, anticipated schedule, and other basic requirements. Data collection specifics concerning data accuracy, routes to be included, route numbering approach, and so on. Division of responsibilities among the agency and the contractor, including a list of the information and assistance that will be provided by the agency. Specifics regarding the agency s data collection requirements and guidelines. Data processing specifics (if any) and data delivery format, including any requirements for the vendor to provide data in a digital user format that can be accessed without proprietary software. Functionalities expected in the software and tools provided by the vendor. Proposed training schedules for agency personnel to collect, process, and manage the data using tools provided by the vendor. Data storage and data hosting responsibilities. Data ownership declarations. B-1

Detailed quality control/quality assurance strategies and timelines for remedies. Historic data integration strategy. Incentives and disincentives based on quality, timeline, and deliverables. Some of the common appendices or addendums included with an RFP are listed below. Selection criteria. Network maps. Data dictionaries for assets to be included in the inventory. Condition assessment manuals (if in the same RFP). A sample contract with standard provisions that will be included. B-2