|This report is an archived publication and may contain dated technical, contact, and link information|
Publication Number: FHWA-HRT-10-073
Date: November 2010
Roadway Geometry and Inventory Trade Study for IntelliDriveSM Applications Applications
CHAPTER 3. ROADWAY GEOMETRY AND INVENTORY DATA SOURCES
This chapter summarizes the findings from the investigation of existing and potential sources of roadway geometry and inventory data and discusses how well the sources satisfy the data needs identified by IntelliDrive stakeholders.
The investigation began with an extensive Web-based search to identify all U.S. commercial roadway database developers, a review of publicly available marketing material and technical documentation provided by those commercial developers, and an email to all State DOTs asking for information about the roadway inventory data they collect and maintain. A series of indepth follow-up interviews were conducted with commercial roadway data developers to obtain additional information about their databases and data collection procedures and to better understand their current business plans with respect to providing additional data needed to support potential IntelliDrive applications. Interviews also were conducted with a selected sample of State DOTs; local highway, public transit, and transportation management agencies; and commercial roadway data collection firms. The purpose of these interviews was to better understand what roadway data are currently being collected by public agencies or what could be collected with emerging technology and to investigate what actions might be needed to encourage public agencies to collect specific data items in support of IntelliDrive applications. Summaries of each of the interviews can be found in appendix C of this report.
The remainder of this chapter is subdivided into three subsections based on the geographical coverage of the data: national, statewide, and local or sub-state. A fourth subsection summarizes the findings regarding emerging data collection technologies.
All of the roadway geometry and inventory data required for IntelliDrive applications are georeferenced, meaning they can be linked to a physical location on the earth and can be located and displayed on a map.
Georeferenced data can be stored and displayed using one of the following four methods:
Each of these methods has implications with respect to the locational accuracy, resolution, and ease of integrating data across sources. To better understand these implications, this chapter presents brief descriptions of each data storage method.
A geospatial roadway network is a vector-based representation of the physical roadway network. Geospatial roadway networks are comprised of two simple geometric objects—points (which specify locations) and lines (which depict the roadway segments that connect two points).
Both points and lines can include attribute data. Point attributes typically include geographic coordinates such as latitude and longitude (and elevation data, where available), which enable the point to be associated with a physical location on the earth. Point attributes can also contain information about the physical location associated with the point, such as a specific intersection or geographic feature type (e.g., bridge).
Line attributes typically contain information about the physical roadway segment depicted by the line, including items such as road name, number of lanes, surface type, traffic flow (i.e., one- or two-way), and grade or slope.
Roadway segments can be of any arbitrary length, depending on the application needs of the network developer. Typically, a new roadway segment is defined whenever there is a change in a key attribute (e.g., number of lanes). Additionally, roadway networks that support vehicle routing applications define new roadway segments wherever two roads physically intersect (e.g., at grade intersections or freeway exit and entrance ramps).
Each roadway segment and endpoint in a geospatial roadway network database has a unique identifier. These unique identifiers can be used to link attribute data stored in other databases to a specific roadway feature for display or analysis purposes. Attribute data that are linked using a feature identifier are assumed to apply to the entire feature. Therefore, if an attribute changes along a roadway segment (e.g., pavement width), the segment must be split wherever the change occurs or the associated pavement width value will be inaccurate over some portion of the segment.
Linking attributes to roadway segments and endpoints using a unique identifier also requires considerable maintenance and updating of the identifiers themselves. For example, whenever a roadway segment is split (e.g., to accommodate a new intersection), the old roadway segment identifier must be retired and three new segment identifiers (plus one new endpoint identifier) must be created. Attributes associated with the old segment must have their segment identifier updated, and new attribute records must be created for each new roadway feature. These changes in feature identifiers must be implemented in all attribute databases linked to the geospatial roadway network using this method.
Some roadway data may be represented as separate geospatial features distinct from the geospatial roadway network itself. Roadway-related data that are typically depicted as separate geospatial features include bridges, railroad crossings, signs and traffic signals, and points of interest (POIs) (e.g., rest areas, airports, etc.).
These geospatial features can be displayed and linked to the geospatial roadway network based on their geographic coordinates. The primary function of GIS technology is to link data stored in different georeferenced databases based on their location information. However, the accuracy with which data stored in different geospatial databases can be linked depends on how accurately the location is measured in each database. The position of a bridge can appear to be off by several meters or more from a roadway feature if location measurements in the two databases were collected based on different accuracy standards.
Only NAVTEQ™ and Google™ currently have transit stop locations and typically only in large cities.
Many roadway attributes maintained by State DOTs are linked to a location on the physical roadway network using an LRS. With an LRS, locations of roadway features such as bridges, guardrails, or signs and changes in roadway attributes such as number of lanes or pavement condition are measured as a linear distance along a defined route from a specific reference point. The reference point could be a physical feature, like an intersection or mile marker signpost, a virtual point, like a State or county boundary, or the starting point of a State numbered route. By specifying the reference point and measured distance, virtually any roadway feature can be located within the accuracy of the distance measurement.
Linear referencing has been used by State DOTs for nearly 40 years and predates the use of GIS technology in most State DOTs by more than 20 years. Indeed, many State DOTs still maintain roadway inventory data on a mainframe computer and have only recently begun migrating these databases into a GIS enterprise environment, which allows the data to be integrated and displayed together with the State DOT's geospatial roadway network.
Keeping an LRS up to date requires considerable effort. New distance measurements must be taken whenever a road is realigned or a route redefined (e.g., a numbered route is switched to another road). State DOTs have developed elaborate internal procedures to update their LRS under different situations, and these procedures vary from State to State and even from one database to another within a State DOT. Additionally, State DOTs develop and maintain an LRS only for those roads for which they have maintenance or administrative responsibility. Most roads within a State have no associated LRS.
In the following discussion of roadway data sources, information is presented on the primary method of data storage and on any alternative methods used. The terms used to describe each of the four data storage methods are geospatial roadway endpoint or roadway segment feature; linked via roadway feature ID; other geospatial feature; and linked via LRS.
The investigation identified six geospatial roadway network databases that provide nationwide coverage and include all, or nearly all, public roads. They are ALK Digital Maps™, DeLorme North America Data Set™, Google Maps™, NAVTEQ™ Road Network Database, Tele Atlas® MultiNet®, and Census Bureau TIGER/Line®. Only one of these databases, the Census Bureau TIGER/Line®, is in the public domain. The other five are proprietary databases, developed for specific commercial applications. Table 5 compares each of the roadway network databases with respect to several key database characteristics. Table 6 through table 9 indicate which of the specific roadway data elements identified by IntelliDrive stakeholders are currently included in each roadway network database.
All five of the commercial roadway networks support vehicle routing, with roadway segment attributes identifying one-way streets, overpasses and underpasses, turn restrictions, numbered route identifiers, and stratification by roadway type (e.g., limited access principal routes versus local streets). Most importantly, the commercial roadway networks segment their roadway features at all at-grade intersections and clearly identify those roadway segments that cross but do not physically intersect (e.g., overpasses and underpasses). This data structure, which is essential for vehicle routing, is defined as network topology.
The Census Bureau TIGER/Line® road network does not support vehicle routing directly. The TIGER/Line® database was developed to help the Census Bureau define geographic areas for the purposes of collecting, compiling, and reporting population statistics. Roads are one of several geospatial features used to delineate Census boundaries. Others include rivers, railroads, and State, county, and municipal borders. All of these features are collapsed into a single layer based on planar topology (i.e., linear segments are defined wherever two linear features cross). There are no explicit attributes or data structures in TIGER/Line® that distinguish between roads that cross at-grade and those that cross but do not physically intersect. Nor is there attribute data in TIGER/Line® that identifies one-way verses two-way road segments or restrictions on turning movements. Any or all of these attributes can be added to the TIGER/Line® geometry. Indeed, most of the commercial roadway databases were originally developed from previous versions of TIGER/Line® road databases. However, these enhancements require considerable time and effort, and the required data must be obtained from other nonproprietary sources, such as State and local transportation agencies.
Only two of the commercial developers, NAVTEQ™ and Tele Atlas®, reported that they are planning to enhance their current roadway network databases in order to support IntelliDrive data requirements. Those enhancements include the following:
While improvements in locational accuracy are being incorporated directly into their core roadway networks, both NAVTEQ™ and Tele Atlas® are developing and marketing their new data attributes as supplemental products to their core roadway navigation database. Measures of roadway curvature and grade are bundled with longitudinal lane markings in an advanced driver assistance system (ADAS) product, while attributes pertaining to pedestrian and transit use are bundled into an Urban Maps™ or Discover Cities™ product.1
Neither NAVTEQ™ nor Tele Atlas® currently provides complete nationwide coverage for all ADAS attributes. In general, attribution is more complete on higher functional class roads (e.g., interstates, other freeways, and major through routes) and in larger urban areas. Attribution is less consistent, or missing entirely, on low-volume rural roads and local streets.
All of the national geospatial roadway networks define roadway segments as the "centerline of the travel way." In practice, this means that roads that allow two-way traffic flow without a center median or physical barrier are depicted as a single line feature. Divided highways that are separated by a median or physical barrier are depicted as separate line features—one for each travel direction. Individual travel lanes are only depicted as a separate line feature when they are separated from the main roadway by a physical barrier (i.e., HOV lanes, exit/entrance ramps, or channelized turn lanes). Otherwise, lanes are represented as attributes of the roadway segment (e.g., number of lanes, average lane width).
NAVTEQ™ and Tele Atlas® are currently collecting data on horizontal curvature, grade, and elevation as components of their enhanced ADAS data products. The data are stored as attributes of individual roadway segments, endpoints, and even of individual vertices (shape points) that are used to define the geometry of a roadway segment. NAVTEQ™ currently calculates roadway horizontal curvature and grade only for those roads where it has collected more accurate roadway centerline coordinate information (i.e., freeways and major through routes). Tele Atlas® reports having horizontal curvature, grade, and elevation for all U.S. roads. However, these measures are calculated from the existing digitized road vectors and are only as accurate as the underlying road database.
None of the national roadway networks currently include measures of vertical curvature, cross slope/superelevation, or available sight distance.
Roadway Inventory and Intersection Characteristics
The geographic locations of roadway intersections, ramps, and most medians (over a specified length) are defined implicitly by the way road segments and endpoints are depicted in a roadway network database. Most at-grade intersections are depicted as a point feature, which represents the approximate geographic center of the intersection itself.
Data on posted speed limits, number of lanes, roadway use restrictions, and restrictions on specific turning movements at intersections are maintained by most of the commercial roadway network developers. More detailed ramp information, lane- and intersection-level attributes, sidewalks, pavement markings, and traffic signal characteristics are being collected only by NAVTEQ™ and Tele Atlas® for specific roadway types and/or geographic areas. None of the commercial roadway networks currently include data on roadway shoulders and guardrails or specific attribute data on medians or individual lanes.
Other Geospatial Features
Most commercial roadway network developers currently maintain data on the locations of highway bridges, tunnels, and railroad crossings. These locations are either stored as a separate geospatial feature (e.g., a point, line, or network) or integrated into the roadway network itself as a roadway segment (e.g., a long bridge) or segment endpoint (e.g., a railroad crossing). In general, these features were originally obtained from public data sources (see next chapter), but the locational accuracy has been updated and corrected to better match the geospatial accuracy of the roadway network database.
Those developers with products oriented toward commercial trucks (including NAVTEQ™, Tele Atlas®, and ALK®) also have attribute data on bridge height and weight restrictions, vehicle restrictions on specific roadways, toll roads, and locations of truck rest areas and weigh stations, primarily on higher functional class roads. Attribute data may have been obtained from public data sources, but the commercial database developers have updated and corrected this information using a variety of sources, including ongoing contacts with State and local roadway agencies, error reporting and feedback from their database clients (e.g., trucking companies), and strategic partnerships with specialized industry data sources.
Only NAVTEQ™ and Google™ currently have transit stop locations and typically only in large cities.
The investigation identified three databases maintained by the U.S. DOT that provide nationwide coverage for some of the roadway attributes identified by IntelliDrive stakeholders. These databases are described in this section.
FHWA's HPMS is a public domain database of roadway condition and performance measures submitted annually by State DOTs in compliance with federally mandated reporting requirements. Several data items identified by IntelliDrive stakeholders are included in the HPMS submission.
HPMS data are used by FHWA to produce its annual Highway Statistics report, which summarizes the current status of the Nation's highways. The data also are used extensively in the analyses of highway system condition, performance, and investment needs that make up the biennial Condition and Performance reports to Congress.
FHWA requires State DOTs to submit certain information about all public roads located within the State. However, data on non-Federal-aid roads (i.e., roads classified as local or minor rural collector) need only be summarized by functional class. Data on Federal-aid roads, including all roads on the National Highway System (NHS) and other roads functionally classified as major collector or higher, are reported by specific roadway segment. A limited number of data items (universe data) must be reported for all roadway segments. These include number of through lanes, number of HOV lanes, average daily traffic, ramp locations, and number of lanes on each ramp. A more extensive list of roadway attributes (sample data) are reported only for a small, statistically selected sample of roadway segments. Sample data include several of the roadway inventory data items identified by IntelliDrive stakeholders such as lane width, median and shoulder types, and speed limit.
The principal value of HPMS as a source of roadway inventory data is that it provides a standardized and consistent set of specific attribute data across all States. However, many of the data items of interest to IntelliDrive stakeholders, including roadway curvature, median, shoulder, and intersection characteristics, are only reported for sample segments in each State. Moreover, some attributes, such as horizontal curve and grade, are reported not as individual occurrences but as summaries or percentages over the entire length of each sample segment.
Another consideration in using HPMS as a data source for IntelliDrive applications is that HPMS data represents roadway attributes along an "inventory" direction. This is particularly relevant on divided highways (i.e., roads with a median or barrier separating directional travel lanes). The data reported in HPMS on divided highways differ depending on the specific data item. For example, reported number of through lanes and average annual daily traffic represent the sum of both directions, while average lane width and pavement condition represent data collected only in the inventory direction.
Lastly, HPMS does not collect segment-specific data for roads that are not part of the Federal-aid system (i.e., local roads and minor rural collectors). These are, in general, the same roads that commercial roadway network database developers find most difficult to accurately attribute and keep current. These roads comprise approximately 75 percent of the total roadway centerline mileage in the United States, although they carry only 15 percent of total annual vehicle miles traveled.(3)
Since about 2000, information on the location of roadway segments reported in HPMS has been stored using a national LRS developed by FHWA. The national LRS allowed FHWA to display HPMS attribute data on the National Highway Planning Network (NHPN), a relatively inaccurate geospatial roadway network representing higher functional class roads, including those that comprise the NHS. However, State DOTs were required to translate roadway data from their own internal LRS to FHWA's national LRS, which often resulted in reporting errors and loss of locational accuracy, since most State roadway networks are more accurate and current than the NHPN. Consequently, beginning with the 2010 HPMS submission, FHWA will require State DOTs to report HPMS segment locations using their own LRS and geospatial roadway network in lieu of the national LRS. State DOTs also must include their geospatial roadway network and LRS as part of their annual HPMS submission so that FHWA can display HPMS data using the State networks.
Web site: http://www.fhwa.dot.gov/Bridge/nbi.htm
NBI is a database containing approximately 600,000 of the Nation's bridges subject to the National Bridge Inspection Standards (NBIS), including all bridges that are open to the public and are longer than 20 ft (6.1 m).(4)
Bridge records in NBI are stored as geospatial point features, independent of any specific geospatial roadway network. Each bridge record contains approximately 100 attributes, including information on the bridge's location, structure and material type, age and service, geometric data, navigational data, classification, condition rating, load rating and posting, and appraisal and inspection types and dates. These data are collected, updated, and maintained by State DOTs and are reported to FHWA on an annual basis in compliance with the NBIS. Both geographic coordinates (i.e., latitude and longitude) and LRS measures are collected for bridges located on Federal-aid highways. The LRS measures are being updated to be consistent with new LRS reporting requirements in HPMS.
The NBI Reporting and Coding Guide allows each State flexibility in determining what point on the bridge is used to define its location (e.g., beginning of span verses midpoint).(5) Currently, NBI does not require States to report locations for bridges located on non-Federal-aid roads.
Although States submit updates to the NBI annually, many bridge attributes, such as vertical clearances and load capacities, are typically updated only after a bridge inspection. NBIS regulations typically require bridges to be inspected at least once every 2 years, but States may request a waiver to extend the inspection interval on certain bridges to 4 years. Consequently, some attribute data in NBI could be more than 4 years out of date.
Web site: http://www.fra.dot.gov/us/content/801
The U.S. DOT's HRCI is a database of all public and private highway-rail crossings, including pedestrian crossings, both at-grade and grade separated (i.e., overpasses and underpasses). Each crossing is identified by a unique crossing inventory number assigned by the Federal Railroad Administration (FRA).
Highway-rail crossing records in HRCI are stored as geospatial point features, independent of any specific geospatial roadway network. Location is stored as geographic coordinates representing the midpoint of the crossing. Each record includes up to 152 attributes, including location, railroad and highway traffic volumes, signal system, and highway and railroad physical characteristics and geometry. These data are collected and maintained jointly by State DOTs and the railroad that owns the track. Previously, there was no regular reporting cycle for HRCI. However, starting October 16, 2010, the Rail Safety Improvement Act of 2008 requires both States and railroads to provide annual updates to HRCI.(6) Additionally, updates to inventory data items are to be reported whenever a physical change occurs at the highway-rail crossing (e.g., installation of a new signal system, significant changes in highway or railroad traffic, addition or abandonment of a crossing track, etc.).
State DOTs in all 50 States and the District of Columbia have developed and currently maintain geospatial roadway networks. The State roadway networks are used primarily to display roadway inventory data collected by the State DOT and to integrate data collected by different parts of the agency based on geographic location. A majority of State DOTs are in the process of establishing agency-wide enterprise GIS data warehouses that will enable agency staff to link any geographically referenced database to the roadway network using geospatial coordinates and/or LRS.
The roadway networks developed by State DOTs vary significantly with respect to locational accuracy, feature representation and resolution, completeness, and network connectivity. Over the past decade, many State DOTs have initiated programs to improve the locational accuracy of their roadway network databases, utilizing high resolution (1:10,000 scale or better) digital orthoimagery or GPS tracks collected by roadway inventory vehicles. The geospatial roadway networks produced from these data collection methods have locational accuracies of ±16 ft (5 m) absolute error or better. According to the most recent annual survey of State DOT GIS coordinators, about 60 percent of State DOTs currently have or are developing geospatial roadway networks at this level of locational accuracy.(7)
Improved locational accuracy typically leads to more detailed representation of roadway features. Most of the roadway networks that have been created using high resolution orthoimagery or GPS also include digitized representations of freeway interchange ramps, separate centerlines for each travel direction on divided highways, and dual centerlines for roads separated by a median or physical roadway barrier.
About half of the roadway network databases developed by State DOTs include all public roads. Other State roadway network databases only include those roads that are maintained by the State DOT or that are eligible for Federal highway funding (i.e., roads functionally classified as major collectors or above). Local roads are often not included or are maintained in a separate roadway database that is used only for display purposes and is updated sporadically.
Because State geospatial roadway networks are used primarily for displaying roadway data and not for network analysis or modeling, many of the State networks have not been thoroughly checked for network connectivity. Therefore, the networks are currently not suitable for vehicle routing or navigation. Additionally, many of these databases do not include key attributes required for vehicle navigation such as one-way streets, turn restrictions at intersections, or clear differentiation of overpasses from at-grade intersections.
A few State DOTs have entered into contractual agreements with commercial roadway database developers (primarily NAVTEQ™ or Tele Atlas®) to use a commercial roadway network as the geospatial roadway network for the State. These agreements typically include provisions for data sharing and statewide licenses that allow various State and local public agencies to use the commercial roadway network without paying additional license fees.
Each State DOT collects and maintains information on the characteristics, condition, and performance of roadways under its jurisdiction. Except for data that are reported to FHWA as part of the annual HPMS submission, there are few, if any, nationwide standards regarding what information is collected, how it is collected, or how it is defined within an agency's roadway inventory database system. Some State DOTs collect and maintain little more than what is required for HPMS submission, while others collect and maintain significantly more data on more roads than required for Federal reporting.
There is also considerable variation in the percent of roads for which State DOTs collect and maintain roadway inventory data. Typically, State DOTs collect and maintain inventory data only on those roads where they have administrative and/or maintenance responsibilities. In most States, State DOTs are responsible for a State highway system composed of interstate, U.S. primary, and State numbered routes (e.g., I 95, US 66, California 1), and the agency delegates responsibility for local roads, collectors, and even some arterials to counties and municipal road agencies. State highway systems typically represent between 10 and 15 percent of all roadway centerline mileage within the State. In a few States, however, the State DOT is responsible for all public roads, except where a local jurisdiction specifically requests administrative responsibility for its roads. In Virginia, for example, the Virginia DOT has administrative and maintenance responsibility for 78 percent of all roadway centerline miles.
Table 10 through table 13 summarize the roadway geometry and inventory data that are currently being collected and maintained by a sample of 16 State DOTs (32 percent of all State DOTs). The information was obtained from responses to an email request sent to all State DOTs and from indepth interviews conducted with State DOTs in California, Florida, Michigan, Virginia, and Washington. The tables distinguish between attributes that the State DOTs collect as part of their internal roadway data collection and those that they are required to collect on selected roadway sample segments for HPMS reporting.
Based on the responses received, none of the State DOTs routinely collect inventory data on non-Federal-aid roads (i.e., local and minor rural collectors). Even Virginia, which has administrative and maintenance responsibilities for many local roads, limits most of its roadway data collection to what it describes as "primary routes," essentially higher functional class roads that include interstate, U.S. primary, and State numbered routes. Data collection on other roads is typically conducted only when needed to support a specific construction or maintenance project.
Information regarding specific roadway geometry and inventory data items collected by State DOTs is summarized in this section.
Approximately half of the State DOTs that responded to the survey and/or interview reported that they are collecting data on horizontal alignment and grade beyond what is required for HPMS sample section reporting. Only about 25 percent of the State DOTs are also collecting data on vertical alignment. Even fewer are collecting data on roadway cross slope, clear zone width, or elevation of specific roadway lanes.
Sight distance is not an attribute that most State DOTs routinely collect as part of their roadway inventory. Sight distance is typically measured as needed on a project-by-project basis to locate roadway pavement markings and signs and is reported as a data element in HPMS only for sample sections and only as a summary measure (i.e., percent of total sample section length that meets the minimum sight distance requirement for passing). Of the State DOTs that responded to the survey, only Michigan DOT reported that they maintain an inventory of passing sight distances for State highway system roads.
Nearly all of the State DOTs that responded indicated that they collect data on the number of lanes, medians, shoulders, ramps, and speed limits as part of their roadway inventory. Data collected includes location, median and shoulder type, widths, and information on special lane use or restrictions (where applicable). State DOTs also collect data on pavement condition, including cracking, rutting, and overall ride quality. Several State DOTs said that they only collect data on paved shoulders, and few indicated that they collect data on special use of ramps or shoulders.
Very few State DOTs collect data on sidewalks or longitudinal pavement markings. Several State DOTs mentioned that pavement markings, guardrails, roadside lighting, and signs are the responsibility of individual State DOT maintenance districts. Inventory data regarding these features may reside within each maintenance district, but the data have not been centralized and, in many cases, may not be available in digital format.
Relatively few State DOTs collect detailed intersection characteristics other than locations derived from their geospatial roadway network and summary information needed to satisfy HPMS reporting. One reason for this is that, in many States, the State-maintained highway system does not extend into urban areas, where many of the more complex intersections are located. Likewise, few if any State DOTs currently have a complete inventory of traffic signals, although several mentioned that they have had internal discussions about developing such an inventory.
Other Geospatial Features
Nearly all of the State DOTs reported that they maintain an inventory of highway-railroad grade crossings and highway bridges principally in response to Federal reporting requirements for NBI and HRCI. None of the State DOTs reported that they maintain a statewide database of transit stops, and only a few mentioned that they keep data on truck related facilities such as weigh stations and truck rest areas. Information is typically limited to those facilities located on interstate and other major routes.
A few State DOTs maintain public traveler advisory Web sites that display information on road and weather conditions, traffic congestion, and highway work zones that could affect travel times and traveler safety. These public Web sites are updated periodically throughout the day, providing information in near-real time.
The quality and availability of geospatial roadway networks at a local level are even more varied than at the State level. While a number of county and city governments throughout the United States have GIS departments and are actively developing enterprise data warehouses for their geospatial data, many others have no GIS capabilities whatsoever. Even among those city and county agencies that are actively developing geospatial data, highest priority seems to be focused on cadastral (i.e., land parcel) data and underground utility networks (e.g., water, sewer, and gas pipelines). Local road networks are usually derived from existing road network databases with little modification or enhancement. The most common sources for local road networks are State DOT road networks (for those States that maintain an all-roads network), Census TIGER/Line® files, and commercial road network databases.
Most large- and medium-sized metropolitan planning organizations (MPOs) maintain roadway networks to support their travel-demand forecasting models. These networks are topologically connected to support network analysis and traffic assignment algorithms, but many of them are simple link-node representations (i.e., roadway segments are depicted as straight lines between two intersections, with no intermediate shape points) and therefore are not positionally accurate. Moreover, most networks used for travel-demand modeling include only higher functional class roads; local roads are not explicitly represented.
The type of roadway inventory data collected by jurisdictions and transportation agencies below the State level can vary depending on available resources, agency staff technical expertise, and local priorities. Even the small sample of local transportation agencies interviewed for this study showed considerable differences.
For example, although Broward and Palm Beach counties are both part of the same urbanized area (Miami, FL), each county collects roadway data according to its own priorities, with little or no apparent coordination across county lines. Broward County is creating a detailed geospatial inventory of all signalized intersection features, including signs, traffic and pedestrian signals, control boxes, and utility lines. Palm Beach County is compiling an inventory of traffic signal locations but with little or no additional data on signal or intersection characteristics. Both counties are considering using a NAVTEQ™ roadway network that was licensed through Florida DOT as their roadway basemap.
Traffic management centers (TMCs), such as TranStar in Houston, TX, generally do not collect new roadway inventory data themselves. Rather, they utilize data that have been collected by the counties and cities that they serve. TMCs with operational traffic management responsibilities (e.g., controlling traffic signals within their service areas) almost always require a database of traffic signal locations and operating characteristics.
Most public transit operating agencies have compiled an inventory of transit stop locations and routes for fixed-route buses operating on local roadways. Transit operating agencies typically build their bus route databases on whatever roadway network is commonly used within their service region, but the agencies may need to add new road segments to display routes that cross private property, such as shopping malls.
Existing roadway data sources at the national, State, and local levels provide a baseline for the type and quality of roadway geometry and inventory data that are currently available to support early IntelliDrive applications. However, there also are several emerging data collection technologies that could reduce the cost of collecting certain data items while substantially improving their quality and positional accuracy. Three of the most promising data collection technologies are described here.
IFSAR, utilized by Intermap Technologies, is an innovative data collection system that uses dual synthetic aperture radar technology mounted on an airborne platform to collect high resolution three-dimensional (3 D) imagery quickly over large geographic areas. IFSAR uses a combination of radar frequency bands (e.g., 300 MHz p band and 8–12 GHz x band) to acquire images of both "first surface" features (e.g., tree canopies) and "bare earth" features. By postprocessing the raw imagery data, Intermap can create both digital surface models showing vegetation and built structures and digital terrain models showing bare earth elevations with all vegetation and structures removed.
By further processing the imagery data using feature-extraction software, Intermap is able to create 3 D road centerline networks with locational accuracy better than 9.84 ft (3 m) horizontal and 3.28 ft (1 m) vertical.
Intermap has collected IFSAR imagery for all of Western Europe and the contiguous United States (CONUS) and has already created 3 D road layers for most of the Western European countries. It is currently seeking customers to support the data processing needed to create a 3 D road network for the United States.
The primary value of this technology is that it would facilitate the creation of a complete all-road network for CONUS with a consistent level of locational accuracy, both horizontally and vertically. Given this network, it would be possible to calculate horizontal and vertical curvature, roadway elevation, and grade for all roads nationwide, including local roads and low-volume rural roads where accurate curvature and grade data are currently missing.
Because the extracted road network represents the approximate centerline of a roadway, IFSAR technology is not able to capture data that would enable computation of cross slope or superelevation along curves. These data would have to be obtained using other technology.
Using a roadway inventory vehicle to collect videologs of roadway features is not new technology. Many State DOTs have been collecting videologs or photologs for more than two decades. Typically, however, videologs have primarily been used to visualize roadside features and surrounding environment, not as a feature extraction source for collecting road inventory data.
Recent improvements in GPS and inertial measurement unit technology combined with more sophisticated feature identification and measurement software make it more practical and efficient to extract roadway features from videologs and to measure their locations relative to the roadway centerline. Commercial roadway data collection services such as Fugro Roadware, Mandli Communications, and Michael Baker currently have such technology available on their vehicles. The primary limitation is the client's (e.g., State and local transportation agencies) willingness to pay the additional costs for processing the videolog images to collect specific inventory data.
Feature extraction and measurement from videologs can be used to collect many of the roadway inventory data items identified by IntelliDrive stakeholders, including locations and widths of travel lanes, medians, shoulders, sidewalks, and crosswalks; pavement markings; locations of signs and traffic signals; locations of guardrails; beginning and ending locations of bridges and tunnels; and virtually any other feature than can be seen from the window of a moving vehicle.
Mobile LIDAR is an optical remote sensing technology that measures the properties of scattered light to find the range of and other information about a distant target. LIDAR can produce highly accurate, survey-grade distance measurements (0.394–3.94 inches (10–100 mm)) both horizontally and vertically relative to the mobile platform on which the LIDAR is mounted. Distances to multiple targets can be measured simultaneously from the same mobile vehicle, delivering highly accurate road inventory measurements with a single vehicle pass.
Mobile LIDAR is currently being deployed by some roadway data collection services, such as Mandli Communications and Michael Baker, as a supplement to videologs to provide more accurate location measurements of roadside features as well as measurement of roadway and roadside contours such as roadway, shoulder, and clear zone widths and slopes; locations and heights of signs and traffic signals; bridge and tunnel clearance heights; and offsets and heights of curbs, guardrails, and median barriers.
Processing of mobile LIDAR point clouds to identify specific features requires significant manual effort, and it currently is being marketed as a cost-effective alternative to in field surveys. However, as data processing software improves, the costs associated with feature extraction and measurement should decrease to the point that mobile LIDAR becomes a practical, cost-effective option for roadway inventory data collection.
Topics: research, safety
Keywords: research, safety, IntelliDrive, Roadway geometry data, Roadway inventory data
TRT Terms: research, Safety and security, Safety, Transportation safety