PERSONNEL

PROJECTS

PROGRAMS

FACULTY RESEARCH GRANTS IN TRANSPORTATION

RESEARCH PROJECT DESCRIPTION

August 24, 2000

Project Title

Integration of 3D Modeling and Neural Networks into a Geographic Information System to Predict Societal and Economic Impacts from Transportation and Urban Patterns.

Project Internet Address

http://www.eng.morgan.edu/~gislab

Principal Investigator, Institution, Telephone Number, Email address

Indranil Goswami Ph.D, P.E.
Assistant Professor, Morgan State University, 443-885-3293, indral@eng.morgan.edu

Co-Principal Investigator, Institution, Telephone Number, Email address

Tony Graham
Doctoral Student, Civil Engineering, 443-885-1442, tgraham@eng.morgan.edu

External Project Contact, Address, Telephone Number

Mr. Glenn Robinson
Morgan State University
National Transportation Center
Montebello D-210A
1700 Cold Spring Lane
Baltimore MD 21251
(443)885-1039
grobinson@moac.morgan.edu

Project Objective

	1.		Develop a model of Baltimore to investigate and predict the effects of urban factors on the development of the infrastructure. The predictions made from by such a model can be useful for Transportation Planning and Decision Making.
	2.		Establish a research thrust in the School of Engineering to create and foster an environment wherein a group of faculty and students will be involved in cutting edge geographic information system (GIS) research. These activities will complement the teaching activities in the GIS and Remote Sensing areas that have already been launched in the School of Engineering.
	3.		Begin activities necessary for the establishment of a GIS Research Center wherein, through a continuous stream of activity, a critical mass of personnel will be involved in GIS research.
	4.		Attract additional funding from state, local and private agencies to expand the scope of activities of the GIS Research Center.
	5.		Produce student research publications in the form of Masters theses and Doctoral dissertations.

Project Abstract

The proposed research presents a strategy for developing a three-dimensional (3D) urban environment simulation system with industry standard computer aided design, traditional two-dimensional (2D) GIS databases, non-traditional 3D databases, animation and simulation techniques and the use of neural network logic input/output theory. The proposed integrated system will provide users a method to predict, simulate and visualize the behavior of 3D models from 2D data, explore the 3D model interactively; and modify the model without changing the databases for exploration and display of alternative physical urban environments. This system acts as a two-way road between the GIS database and the predictive model.

The Neural Network Model will serve as a predictive tool for the temporal fluctuations in the state of an urban system. The network will be ‘trained’ using historical data gathered by city, local and state agencies. The dynamic interactions between neighborhoods will be an integral part of the model. The output from the model can be used as a planning tool, following extensive calibrations.

Task Descriptions

	1.		Project coordination and report preparation
	2.		Software training – Phase I
	3.		Software training – Phase II
	4.		Data review and assessment
	5.		Artificial neural network research
	6.		Urban model development, refinement and testing
	7.		Development of customized GUI for project

Milestones, Dates:   

	July 1, 2000		Start Date
	July 30, 2000		Install all hardware, software. Complete Phase I of software training. Setup functioning website and bulletin board. Define individual project goals and objectives.
	August 15, 2000		Begin Phase II of Software Training. Define individual student project tasks. Identify data sources for each project. Prepare executive summary for project.
	September 1, 2000		Start background research on artificial neural networks
	October 15, 2000		First Quarterly Report
	October 30, 2000		Complete background research on artificial neural networks and Bayesian belief networks. Start developing urban model. Collect physical, demographic and socio-economic data for Baltimore Metropolitan Area.
	November 30, 2000		Refine urban model.
	December 31, 2000		Test urban model using historical data for Baltimore.
	January 15, 2001		Second Quarterly Report
	March 31, 2001		Refine urban model.
	April 15, 2001		Third Quarterly Report
	June 30, 2001		End Date

Yearly and Total Budget

Budget for July 2000-June 2001: $ 84,994

Student Involvement (e.g. Thesis, Assistantships, Paid Employment)

	1.		Doctoral Student/Civil Engineering – Will pursue research activities that will serve as dissertation research.
	2.		Senior Student/Civil Engineering – Will handle webmastering services and Use of AML (Avenue) for Integrating User-Scripts into GIS Functionality.
	3.		Masters Student/City and Regional Planning – Will perform some training functions and pursue a research project, which may develop into a masters thesis.

Relationship to Other Research Projects

GIS activities in Dr. Fred Wilson’s research group (also in the School of Engineering) will eventually have significant coherence with activities sponsored through this grant. Within this grant period (July 2000 – June 2001), collaboration possibilities between these two groups will be explored. 

Technology Transfer Activities

Development of a database for Baltimore City Urban Network is a primary objective of this project. City planners and transportation planners can use the database and the associated model.

Potential benefits of the Project

	1.		Development of an urban model of Baltimore
	2.		Establishment of the GIS Research Center as a new capability of the School of Engineering
	3.		Attract additional funding in the form of sponsored projects
	4.		Training of students – theses and dissertations

TRB Keywords:

Planning, Neural Networks, Geographic Information Systems, Urban Modeling, Baltimore, Maryland, Urban Data, ArcView, ArcINFO