GIS Traffic Planning Tools

Project Statement

The objective of this project is to develop tools and procedures to assist Iowa transportation planners and traffic engineers in conducting various impact studies. The Iowa DOT Transportation Planning and Modeling Research Needs Focus Group has identified three types of studies, in particular, that are likely to benefit from improved modeling techniques. These are: site impact analysis, interchange and other major infrastructure justification and bypass analysis. Recently developed computing technology can be applied to these problems (e.g., geographic information systems (GIS), windows-based travel modeling software and internet-based information access methods).

Background

Where available, current Iowa urban transportation planning (UTP) models are oriented to regional, 24 hour modeling and forecasting. These models provide, in general, adequate estimates of current and future average daily traffic on a region-wide basis or for major flows such as through freeway interchanges or along principal corridors. However, for studies related to localized project development (road projects or land development), it is more important to estimate traffic during periods of peak flow. Further, where models are not available (e.g. for nearly all Iowa places under 50,000 population), current traffic models do not even exist. For many of these locations, studies of the traffic impact of road and development projects must be conducted without the aid of modeling technique. In many of these studies, such as when analyzing the impact of a proposed town bypass, a travel model would provide reliable traffic estimates, useful in comparing alternatives and assessing benefits and costs.

Since the 1960s, transportation modeling efforts have evolved through several phases. In the 60s and early 70s, interest and funding was available to support the development of mainframe urban travel models in many areas. Large surveys were conducted and significant efforts resulted in the proliferation of models in Iowa. Models were developed for cities as small as 25,000 by the DOT. In the late 1970s, funding became more restricted and federal legislation mandated modeling only in areas greater than 50,000 (metropolitan planning organizations, MPOs). Quick response planning techniques (e.g., NCHRP 187) were developed so that model parameters could be transferred from existing models so that surveys could be smaller or not conducted at all. Personal computer implementations of travel models produced descendants of the mainframe techniques (e.g., Tranplan) and a PC based quick response system (QRS). A statewide public sector Tranplan license was purchased by the DOT in the early 1990s, and five of the eight Iowa MPOs and one city (Ames) use this model. The other three MPOs use the latest version of QRS (QRSII). Tranplan has found a niche as a powerful, comprehensive modeling package, while QRS is noted for user friendliness and excellent site impact and traffic engineering capabilities.

During the same period, GIS software has evolved and is beginning to be used by state DOTs and local transportation agencies. The superior data management, query and presentation capabilities of GIS have been integrated with the travel forecasting models in a previous Iowa DOT/CTRE project. In fact, a procedure for developing travel models for smaller urban areas has already been developed which uses Tranplan and MapInfo GIS. This system has been used effectively to develop a travel model for the City of Ames. A current CTRE/FHWA project is porting this system into other desktop GIS platforms and is assessing the usefulness of the tool to state and local planning.

There are several limitations of the existing system which inhibit its direct application to the three types of studies described above. First, the existing system is based on Tranplan whereas QRSII is likely to provide improved local impact capabilities. Second, the procedure in its current form develops only 24 hour models whereas peak hour models would be much more useful. Third, the system has only been tested and implemented in areas where model developers had extensive local knowledge and assistance. In many of Iowa's smaller urban areas, local expertise does not exist.

As mentioned, several of the Iowa MPOs use Tranplan, others use QRSII. For the reasons discussed above, Tranplan users may wish to access QRS capabilities for their areas, and visa-versa. However, developing new models in the alternative platforms is an expensive option. While developing the Tranplan Mapinfo system, CTRE researchers have also developed a limited link between QRSII and Mapinfo. It is possible that with modification, the two systems could be used to automatically translate existing Tranplan models into QRS and existing QRS models into Tranplan.

Further, problems are known to exist with commercially available, "sequential" travel models (like Tranplan and QRS). These problems relate to the potential of non-convergence of traffic assignments in congested networks and the lack of the ability of the models to represent user behavior. These problems are not likely to be important limitations to network modeling in Iowa due to relatively uncongested conditions. Two alternative approaches (simultaneous modeling and dynamic modeling) have been proposed, but to date, are not readily available to practitioners. These models are still likely to be years away from practical application, but it is possible to evaluate the tools developed in this project for the non-convergence problem. This effort would have the added advantage of allowing for the explicit incorporation of intersection performance in the travel model (currently QRS allows this to some extent, Tranplan does not). Model turning movement forecasts can be fed back into a representation of intersection turn penalties using the GIS system.

Finally, there has been a reluctance on the part of the traffic engineering community to rely upon site specific forecasts of traffic from UTP models. This is in part due to the purpose of the sequential models, which was never supposed to provide pinpoint estimates of link specific flows. However, methods of extrapolation and expert judgment to forecast flows also leave something to be desired. For example, summing extrapolated traffic estimates for a region would likely produce unrealistic region-wide totals. With the proposed tools, especially considering the feedback of turning volumes representing intersection delay, travel estimates produced by UTP models might be more useful to traffic engineers.

Literature Review

A search of TRIS and other reference sources was performed, (see Section IX, Information Sources and References for a selected bibliography). Some of the most recent work done nationally on integration of travel models and GIS has been performed by CTRE for the Iowa DOT. Two research projects, three masters theses, and several publications and conference presentations have resulted from this work. An NCHRP Research Results Digest (number 210) presents a literature review of the effects of highway bypasses on small communities, referring to much of the recent work in this area.

Proposed Research Methodology

We propose that the development of GIS Traffic Planning Tools be staged into five sequential tasks, as follows:

(5%): GIS Traffic Planning Tools Research (2 months)

Review Literature and State of the Art: While national research efforts are leading toward the development of "usable" simultaneous and even dynamic-based travel forecasting models, these models are not likely to be developed within the time frame of this study, and may be several years off in terms of practical implementation. Therefore, while these efforts will be reviewed for possible inclusion of techniques into the GIS Traffic Planning Tools design, QRSII and Tranplan (the two predominantly used travel forecasting packages used in Iowa) will form the basis of the tools developed in this project. New NCHRP (365) guidelines are also coming out, and will be reviewed for inclusion in tools development. References for development of peak hour models will also be reviewed.

Review Available Software: Several GIS and access platforms will be investigated (e.g., ArcView, Maptitude, Mapinfo, internet application software, etc.) Preference will be given to software that would allow free or low cost user access (e.g., public domain GIS, or possibly remote access application via the Internet).

(30%): GIS Traffic Planning Tools Design (3 months)

Design QRS/GIS Environment: Design will incorporate existing Tranplan/GIS functionality including import/export, visualization and calibration tools.

Devise method for refining 24 hour models to provide peak hour estimates: The method will be implementable in Tranplan and QRSII.

Define procedures for developing models for small urban areas without local assistance: Additional quick response procedures adapted from NCHRP 365 and previous work will be incorporated into the modeling framework to allow efficient development of calibrated models for smaller urbanized areas without full time traffic planning staff.

Design Tranplan/QRS translator: An efficient method for converting Tranplan models to QRS and visa-versa will be devloped. Due to size limitations for various versions of QRSII, some models may not be readily converted from Tranplan to QRSII.

Design method for estimating intersection parameters based on flow and traffic engineering principals: The Tranplan/GIS interface will be enhanced to allow for the explicit inclusion of volume-derived intersection delay.

For all subtasks, the project team will assure that software and databases maintain compatibility with existing and planned DOT computing standards and readily available hardware/software. The team will inform the DOT GIS coordinating committee of progress and request their review of same.

(40%) GIS Traffic Planning Tools Development (4 months)

i) program systems in GIS and UTP model environments

ii) acquire data to test GIS based traffic planning tools

iii) develop access/distribution methods

iv) prepare users manuals and documentation

(15%) GIS Traffic Planning Tools Testing/Refinement (2 months)

i) identify beta test sites for each planning tool (with input from Project Advisory Committee)

ii) test systems at beta sites

iii) refine systems based on feedback and PAC input

(10%) Technology Transfer (1 month)

i ) prepare project report

ii) deliver software and documentation

iii) GIS Traffic Planning Tools training

Project Products and Deliverables

For each of the systems described in Section IV. C. (above), deliver component software and users manuals to the Office of Systems Planning and interested MPOs. Provide suitable training for Systems Planning staff and interested MPO staff (at Iowa DOT or CTRE).

Project Personnel

Principal investigator: Reg Souleyrette, Associate Professor of Civil and Construction Engineering at ISU. Reg is also the Associate Director for Transportation Planning and Information Systems at CTRE. Reg will be responsible for project management, coordination including meetings and needs assessment.

Research Associate: Mike Anderson, Department of Civil and Construction Engineering, ISU. Mike wrote his masters thesis on integration of GIS and transportation planning models. He will be responsible for coding, data collection, documentation, testing and training for the GIS Traffic Planning Tools.

GIS Specialist: Zachary Hans, CTRE. Zach will provide guidance on GIS design and assure that GIS Traffic Planning Tools are compatible with related DOT efforts.

Information Sources and References

Potentially Useful References:

 

CTRE is an Iowa State University center.

Address: 2711 S. Loop Drive, Suite 4700, Ames, IA 50010-8664

Phone: 515-294-8103
FAX: 515-294-0467

CTRE Communications: Marcia Brink
CTRE Webmaster: Michele Regenold

Website: www.ctre.iastate.edu/

Iowa State University--Becoming the Best