He’s Driven to Predict Traffic Jams

Just as meteorologists can forecast the weather days in advance, a Georgia Tech researcher is designing a system that can predict traffic jams long before rush hour begins.

The software being developed by Georgia Tech civil engineering professor John Leonard will predict traffic flows or “temperatures,” showing drive times in bright colors like a weather map. He also is developing a scale for gridlock so traffic conditions can be expressed in a figure.

“People need a simple-to-understand number--even if it doesn’t have a physical meaning--to represent traffic congestion,” Leonard said. “What I want to do is bring traffic-flow science to the traveling public . . . and get people to start developing a personal understanding of congestion and plan accordingly.”

Just as people use weather reports to plan how they will dress the next day, they could use traffic reports to plan their commute.

“You know at certain times of the day it’s going to be bad no matter what,” said Marion Jones, state traffic operations engineer for the Georgia Department of Transportation. “But this might be able to tell you if you should leave 15 minutes early or can hit the snooze.”

Leonard’s current drive-time predictor helps with an immediate departure. It uses historical data collected from the state DOT along with variables such as weather, population-growth figures, the calendar and events such as ballgames or major conventions.

Current users can only click on five different departure points on the map--Hartsfield Atlanta International Airport, downtown and the main junctions of the interstate system around the city--to get drive-time estimates.

But by March Leonard says someone will be able to click anywhere on the map and make that the departure point, so someone could check the estimated travel time into the office before leaving home.

He also says predictions made 24 hours in advance could be online early next year, allowing drivers to create custom drive-time estimates.

“That would be the big-picture stuff,” Leonard said. “Then, as we get closer to real time, we could fine-tune it with what’s going on now.”

Leonard said the concept can be applied to any major city with a traffic-management system that tracks flow and average speeds and has the historical data necessary for predictions.

“Most major metros have something up and running,” Leonard said. “It’s just that Georgia’s and Atlanta’s is one of the most extensive and offers us a chance to show the future of this kind of stuff.”

*

The Web site is https://traffic.ce.gatech.edu/trafficweather.