DG Project Makes Traffic the Network
Georgia Tech Team Uses Car-to-Car Communications to Model a Clear Route Through Traffic Congestion
By Karen Heyman
For the DGRC

Modeling Better Traffic
	Researchers • Richard Fujimoto • Randall Guensler • Michael Hunter Project profile ITR: Simulation-Based Operations Planning for Regional Transportation Systems Home page r

It often feels as though life's inevitabilities should include, "Death, taxes and roadway congestion."

Many of us have come to accept that it's just one of those things you cannot do anything about except complain. But in Atlanta, which like any major metropolitan area, suffers from too many cars trying to get to too many places at the same time, local officials and Digital Government researchers are trying some innovative ideas to keep traffic moving.

The current solution, as it is in many major cities, is to deploy hundreds of cameras and other detection devices to monitor road conditions and report back to central offices. There decisions can be made about traffic alerts and detours. The system is efficient, but expensive, with much of the cost going to maintenance and operation of the system, especially the detection devices.

Atlanta officials are already experimenting with alternatives. The Federal Highway Administration and Georgia Department of Transportation funded the deployment of 500 vehicles in the Atlanta region capable of collecting and transmitting second-by-second vehicle activity data. This instrumented vehicle fleet has been on Atlanta roads since last August, sending back detailed data for hundreds of thousands of vehicle trips.

Digital Government researchers at Georgia Institute of Technology are building on the results of that real-world test to model a potentially even more cost-efficient system.

The current data collection project - being conducted by Georgia Tech faculty members Richard Fujimoto, Professor, College of Computing, Randall Guensler, Associate Professor, School of Civil and Environmental Engineering, and Michael Hunter, Assistant Professor, School of Environmental Engineering - essentially follows the client/server computing model:

An individual car is equipped to communicate wirelessly with a central office. All the cars can send and receive information to and from the office. Still, projects that sound great on paper and work well as demonstration projects can have severe limitations when scaled up to a metropolitan area. In this case, potential solutions to traffic congestion could experience congestion of their own - too many cars reporting at once may clog the communications network, leading to failures of the traffic management system.

The Georgia Tech researchers' plan is to create a system featuring vehicle-to-vehicle communications, where information can flow from car to car as well as from the traffic stream to decentralized operations centers. Vehicle-to-vehicle communications systems could provide many more paths for information to flow through the transportation network. The hope is that in such a distributed system essential information would move quickly and cheaply, leading to a system that is less costly to maintain.

To test the idea safely, Fujimoto, Guensler, and Hunter are working together to create a computer simulation to model real-world traffic and vehicle-to-vehicle communications conditions. They are using for their inputs traffic flow and roadway infrastructure data obtained from their government agency partners, who have been generous with help. The City of Atlanta, Fulton and Cobb Counties, Atlanta Regional Commission, Georgia Department of Transportation and Georgia Regional Transportation Authority have all provided data for model development. The Georgia Department of Transportation has even donated old traffic equipment. Additionally, they are using USGS satellite imagery - which allows traffic planners to remotely determine roadway characteristics, such as how many cars can stack up in a left turn lane.

Although the USGS isn't a direct partner, the images have been so useful, says Guensler, "They should get tons of kudos for what they've done."

The researchers are using an approach to simulation that focuses on software reuse by integrating complex simulations into distributed computing environments, rather than reinventing models from scratch. Currently, the simulation testbed employs CORSIM, a standard traffic simulation program, and QUALNET, a commercial software package used to simulate the vehicle-to-vehicle communications layer. One of the key components of the modeling regime is tracking individual vehicles throughout the system, by the use of unique vehicle identifiers to simulate the peer-to-peer communications. The current traffic simulation cannot scale up to the complexity of the entire city of Atlanta, so the prototype focuses on a demonstration of the communications systems in the heavily populated Northeast quadrant of the region. However, when the time comes, the work can be transferred into more robust modeling programs.

Guensler and Hunter in Civil Engineering are concentrating on traffic simulation model development and the iterative process of checking data inputs against field data, while Fujimoto's team is concentrating on creating the software architecture of the program and simulating the communications network.

Where once it may have been enough simply to write an enormous program and set it to run on a mainframe computer, Fujimoto says, the new challenge for computer experts is to design programs that run efficiently in distributed computing environments composed of personal computers connected through a local area network - a configuration commonly found in offices and labs everywhere. The researchers are also adapting their software to run in a scheme called grid computing, the goal of which is to make computing as ubiquitous and as easy to use as plugging an appliance into a wall outlet without concern for where the programs are executed, or what computation resources they need. The research team is near completion of the test bed and will soon be exploring how traffic control centers and even intelligent traffic control devices can benefit from the data collected and transmitted by the system.

For all the potential savings in time and frustration that the system should eventually provide, use of vehicle identifiers leads to inevitable questions about data privacy. The demonstration model assumes that not every car will need to be equipped to collect and transmit data; one of the questions being examined is what is the least number of cars needed to provide the optimum amount of data.

Latest DG News

• dg.o 2006 Convenes May 21-24, 2006   • dg.o 2006 Early Registration Ends April 10th! • dg.o 2006 Issues CFP - Tutorials • dg.o 2006 Issues CFP - Workshops • dg.o 2006 features Workshops on:    eRulemaking    GeoInformatics • dg.o 2006 features Tutorial on:    •Social Network Analysis • New DG Team Pursues eRulemaking • IEEE ISI2006 Convenes May 22-24, 2006 • eChallenges e-2006 Issues CFP • DG Research Helps Predict Urban Growth • Swapping Secrets of the Double Helix • UK and DO-Wire Launch e-Gov Best Practices wiki • DG Team Develops "Virtual Agora" for e-Gov • Mapping for Times of Crisis • Exploring Detection of Crisis Hotspots • Report: Mass eMail Campaigns Harmful • Scenario-Based Designs for Stat Studies • US, EU Explore Info Integration • DG Team Develops Digital Interpreter • DG Study Gives Teeth to FBI • Research Smooths Road for Small Businesses • DG Researchers Parsing in Tongues • e-Gov Journal Issus Call for Articles • See all news stories Contribute to dgOnline

Even so, with more and more car manufacturers installing GPS systems and many states offering toll transponders, "There will eventually have to be a public accounting to decide what level of information should stay in the public and private domain," says Guensler, who is writing a paper that discusses privacy issues.

But in the meantime, he points out that the potential benefits from instrumented vehicle data and complex simulations are enormous. Traffic data can be used by transportation planners and engineers in real time to target roadway improvements, adjust traffic signal timing, reduce traffic delay, improve safety, reduce fuel consumption, and generally operate the transportation system in a much more cost-effective manner. "Keeping a smooth traffic flow, without sudden accelerations, is also optimal for reducing pollution," says Guensler.

Commenting on the value of the Georgia Tech team's work, government partner Shaun Green, a Transportation Engineer with the Georgia Regional Transportation Authority, says, "Georgia Governor Sonny Perdue put together a traffic task force to improve traffic signals, basically to make sure they're not the primary cause of congestion. There's a group of 40 or 50 people in thirty different jurisdictions who work on it. Michael presented an overview of the research to us. Nobody knew what to expect, and everybody left thinking, wow, this is cool - and these are people who do this every day for a living, and always feel they don't have enough time or money for everything that needs to be done. They were all thrilled to death with the potential results of the research."