Beyond ‘Real Time’ Reporting, Forecasting Future Traffic Jams - NYTimes.com
With huge increases in the use of traffic cameras, roadway sensors and GPS-linked devices like cellphones, transportation agencies are now grappling with a flood of data available to tell them how well vehicles are (or are not) moving.
Collected and crunched by highway authorities and private companies, the data can be shared with travelers through radio alerts, dashboard navigation systems, Web sites and electronic road signs. “Real time” is the phrase often used to convey the urgency of this information stream.
Trouble is, in many cases real-time traffic updates are just not fast enough.
I.B.M. is teaming up with the New Jersey Turnpike Authority to tackle this problem. The Smart Transportation initiative will use I.B.M. software to predict traffic patterns up to an hour into the future on state roads, including the New Jersey Turnpike and Garden State Parkway.
A series of complex algorithms and analytics make up the software tool that I.B.M. has engineered to instantly interpret the millions of bits of data that flow into a traffic management center in Woodbridge, N.J. But instead of just alerting drivers to accidents, bottlenecks and dangerous conditions, the software will estimate what traffic will be like in 10, 20 or more minutes, helping drivers to make choices.