System Could Predict Traffic Jams Before They Happen

April 14, 2011
Collaborative effort aims to analyze real-time traffic patterns to forecast faster and safer routes

Computer giant IBM Corp. has joined with the California Department of Transportation (Caltrans) and California Center for Innovative Transportation (CCIT), a research institute at the University of California, Berkeley, to develop an intelligent transportation solution that will help commuters avoid congestion and enable transportation agencies to better understand, predict and manage traffic flow.

IBM Research has developed a new predictive modeling tool that will allow drivers to access personalized travel recommendations designed to help them avoid congestion, and save time and fuel.

By joining forces, IBM, Caltrans and the Mobile Millennium team within CCIT hope to provide drivers with predictive information on what traffic patterns are likely to look like – even before they leave work or home and get in their vehicles – rather than discover what has already happened and is being reported.

“As the number of cars and drivers in the [San Francisco] Bay Area continue to grow, so too has road traffic. However, it’s unrealistic to think we can solve this congestion problem simply by adding more lanes to roadways, so we need to proactively address these problems before they pile up,” says Greg Larson, chief of the Office of Traffic Operations Research, Caltrans.

Traffic delays caused by highway incidents such as work zones, crashes or simply by morning and evening rush hours routinely stymie frustrated drivers. Even with advances in GPS navigation, real-time traffic alerts and mapping, daily commute times are often unreliable, and relevant updates on how to avoid congestion often reach commuters when they are already stuck in traffic and it is too late to change course. This inability to avoid traffic congestion has led to commuters across the United States wasting an average of a week's worth of time, 28 gallons of gas and $808 over the course of a year.

In Silicon Valley, the problem is especially acute. In comparison with cities of a similar size in population, drivers in the city of San Jose waste a cumulative of 10 million more annual hours sitting in traffic jams and suffer a 15% higher commute delay per peak-time traveler.

Spanning the San Francisco Bay Area Region, the new Smarter Traveler Research Initiative collects and analyzes traffic data generated from existing sensors in roads, toll booths, bridges and intersections. This project combines that data with locations based on GPS sensors in participating people’s cell phones to learn their preferred travel days and routes. Alerts are then automatically delivered via email or text message on the status of the driver’s typical commute before the trip begins, which eliminates potential driver distraction once the trip is in progress.

These alerts will enable drivers to plan and share alternative travel routes, improve traveler safety and help transportation authorities better predict and reduce bumper-to-bumper traffic before congestion occurs through improved traffic signal timing, ramp metering and route planning.

The researchers will leverage a learning and predictive analytics tool called the IBM Traffic Prediction Tool (TPT), developed by IBM Research, which continuously analyzes congestion data, commuter locations and expected travel start times throughout a metropolitan region that can affect commuters on highways, rail-lines and urban roads. Through this Smarter Traveler Research Initiative, scientists could eventually recommend better ways to get to a destination, including directions to a nearby mass transit station, whether a train is predicted to be on time and whether parking is predicted to be available at the station.

“In order for intelligent transportation systems to be truly effective, travelers need information they can act upon before getting stuck in traffic,” says Stefan Nusser, functional manager, Almaden Services Research, IBM. “By actively capturing and analyzing the massive amount of data already being collected, we’re blending the automated learning of travel routes with traffic prediction of those routes to create useful information that focuses on providing timely, actionable information to the traveler.”

Related Articles

U.S. DOT Issues Challenge to Do Something About the Weather

Everybody talks about the weather but nobody ever does anything about it, as the saying goes, but apparently the U.S. Government thinks all it takes is a little bit of financial incentive to get somebody to do something about it...

Latest from Transportation & Distribution

96378710 © Nattapong Boonchuenchom | Dreamstime.com
#53673151@Petar Dojkic|Dreamstime
Trucking Industry Objects to DOL Rule on Contractors
Trucking Industry Objects to DOL Rule on Contractors
Trucking Industry Objects to DOL Rule on Contractors
Trucking Industry Objects to DOL Rule on Contractors
Trucking Industry Objects to DOL Rule on Contractors