Scott Higgins

Comprehensive Pavement Management in Georgia Utilizing the Route-System Data Model

The Georgia Department of Transportation's Planning Data Services Bureau contracted with the University of Georgia's Government Information Services Division to develop statewide basemap layers to support GIS applications in the area of pavement management. One of the major requirements of this initiative was to establish a relationship between the basemap's route features and the Department's Road Characteristics Database (RCFILE). The RCFILE contains attribute information (linear, continuous, and point events) for state, county, and city level routes and is based on a "county-route-mile point" linear referencing system. The resulting basemap contains thousands of ArcInfo route features for each county which have been calibrated and quality checked using attribute information pulled from the RCFILE.

This paper discusses specific database design issues related to the Georgia DOT's Pavement Management System. Topics discussed include (1) the Route-System Data Model, (2) using dynamic segmentation, (3) working with route features, (4) the RCFILE, and (5) supported pavement management applications.

Background

The Georgia Department of Transportation's Planning Data Services Bureau entered into a services contract with the University of Georgia's Government Information Services Division in June 1992 to develop statewide digital basemap layers to support GIS applications in the area of pavement management. One of the major requirements of this initiative was to establish a relationship between basemap route features and the Department's Road Characteristics Database (RCFILE). The resulting basemap contains thousands of ArcInfo route features for each county which have been calibrated and quality checked using route attribute information contained in the RCFILE. The technical data processing aspects on such an endeavor go on endlessly, however, this paper will discuss broader points of the overall GIS database design. Anyone who is interested in taking a closer look at technical aspects of this project should refer to the 1996 Esri Conference Proceedings, paper #'s 170 and 141.

The Road Characteristics Database (RCFILE)

The Road Characteristics Database, in its current digital form, has been under continuous development by the Department since 1975, serving as the first linear referencing system designed for the statewide collection and storage of route attributes. The RCFILE contains attribute information (linear, continuous, and point events) for 110,000 miles of public roads maintained by the state, including 18,000 miles of the Georgia state highway system. Route attribute event information is recorded and stored within a "county-route-mile point" based linear referencing system. Within this system, event locations which occur along each route are recorded from field inventories to the nearest 1/100th of a mile (.01 miles). RCFILE information is stored by county for each of Georgia's 159 counties and is complete for all Department inventoried routes. Inventories are collected by Department District Offices on a 5 year cycle, and the field inventory process is currently being updated by the Department to include Global Positioning Systems (GPS) technology and database applications to automate the entry of RCFILE data. Once integrated into the Department's data processing flow, the automated inventory collection method should prove to be more efficient and accurate, both relationally and spatially, than previously employed manual collection methods. The Department is also in the process of migrating the RCFILE from a flat ASCII file format to a relational database management system (RDBMS). The following table contains merely a portion of the attribute items contained in the RCFILE which can be mapped using dynamic segmentation:

Route Layer Database Design

One of the Department's primary system requirements was a relational link between the GIS basemap and the RCFILE. An accurate link to the linear referencing system and its attribute data was necessary in order to utilize the capabilities of dynamic segmentation. It was determined that a concatenation of the first 3 attribute fields of the RCFILE constituted a relational key item which would be unique for all route features statewide. In the initial route layer arc-node digitization, performed using 31,680 scale mylar source maps, this field (RCLINK) was captured as an arc attribute for all arcs which composed a given route, along with a point feature denoting the route's zero (0) mile point. Thus capturing both the relational link required, as well as the spatial beginning of the route.

While many other state transportation agencies have created 1:24,000 scale basemap layers for road features, the Department had many reasons for converting its own 1:31,680 scale maps. The source maps, which were originally derived from USGS 7.5 minute quadrangles, had been maintained by the Department for years as new roads were added and other RCFILE attributes were collected in the field. Georgia is fortunate to have statewide Digital Orthophoto Quarter Quadrangle (DOQQ) coverage currently under production, which will enable the future correction of 1:31,680 scale basemap geometry to meet National Map Accuracy Standards (NMAS) for 1:12,000 scale mapping.

Once the GIS basemap layers were completed for all Georgia counties, the next step in the process focused on the building and calibration of route features within the route layer. An AML/C solution was developed initially to automate the creation and calibration of ArcInfo route features. This process isolated the spatial beginning of the route, created the route feature, and calibrated it at its beginning and ending with real world values (field collected route lengths). However, it was quickly discovered that in order to dynamically segment RCFILE events accurately, further route calibration would be necessary. For instance, linear events which occur on a given route and end at the intersection of another route may fall short or extend past that intersection. This is caused by attempting to dynamically segment events referenced by ground measurements on top of a generalized cartographic representation of a route. In order to overcome these types of problems, an intersection level calibration was performed for all routes in the system using a process of isolating X,Y coordinates in the arc-node geometry and intersection mile points contained in the description field of the RCFILE. This process resulted in a calibration point coverage which was then used to calibrate routes. To put it differently, each route in the system was calibrated at its beginning (0), at its intersection with all other routes, and at its end. The ArcInfo route system data model supports this high level of calibration. However, the automation of this level of route development was largely an "in house" accomplishment, requiring large amounts of manual editing, data entry, and processing time.

Supported GIS Applications in Pavement Management

The Department required, like many other organizations, that cartographic applications would be among the first developed to utilize the GIS database. Throughout the initial data conversion effort other ancillary layers of mapping information were captured statewide including political boundaries, railroads, transmission lines, wetlands, lakes, linear hydrology, and cultural information. Several other layers of data still needed to be developed in order for the Department to have all of the required cartographic elements present in the GIS basemap. The entire statewide basemap is currently in a revision cycle which involves the development of additional cartographic layers along with the update of route and hydrology layers. AML software has been developed to produce the Departments County Map Series in a semi-automated cartographic process. New county and city maps are now being produced as basemap revisions are completed for each county.

The Department has also used GIS to produce specialized reports to the Federal Highway Administration (FHWA) detailing the linear referencing system for the Georgia State Highway System, the U.S. Highway System, and other major arterial routes. Other application areas are currently being realized by different groups within the Department as internal GIS knowledge and awareness spreads. Applications in pavement management which deal with analysis of overlapping events, inventory management, reconstruction planning, and specialized mapping are envisioned for the future.

Conclusion

The Georgia Department of Transportation, in cooperation with the University of Georgia, has taken unique approaches to statewide GIS data collection and development. In just four years, the Department has built statewide digital basemap layers, updated its cartographic function from manual to semi-automated methodologies, and updated its field data collection function to include GPS and other data automation technologies. The Department is currently training field crews in new collection methods in anticipation that all new field data collection and integration will be performed with respect to, and integrated with, the Department's overall information processing flow.

Scott Higgins, GIS Project Leader
Carl Vinson Institute of Government
University of Georgia
201 N. Milledge Avenue
Athens, GA 30602-5482
Telephone: (706) 369-6062
Fax: (706) 542-9301
E-mail: scott@lislab.uga.edu