A new system in San Francisco learns from the past to predict traffic jams before they happen
IT鈥橲 the bane of drivers the world over. You round a corner on a busy highway to find the traffic in front of you at a standstill, blocking the road ahead as far as the eye can see. What should have been a quick commute has just turned into a frustrating slog.
For a few lucky drivers in San Francisco, this should be a thing of the past. They are now able to peer into the future and receive warnings of traffic jams before they occur, allowing them to change their route or timing.
Advertisement
The Smarter Traveler Research Initiative blends real-time traffic data with past traffic patterns to predict congestion up to 40 minutes into the future. Drivers are then automatically sent an email or text message of conditions on their regular commute before their trip begins.
The initiative is a joint project by the California Center for Innovative Transportation (CCIT) at the University of California, Berkeley, the California Department of Transportation (Caltrans) and IBM.
鈥淚f you are already on the road and a sign says 鈥榗ongestion next 5 miles鈥 you may have very few options,鈥 says Naveen Lamba of IBM. 鈥淏ut if you get that information prior to starting your journey, you can choose to stay at home, work late or take a different route.鈥
A number of companies, including Microsoft and Google, already provide real-time traffic information and trends of traffic patterns. The California initiative goes a step further to provide detailed information on what traffic will look like in the near future, and personalises it for each individual commuter鈥檚 journey.
Much of the congestion on streets and highways is caused by accidents that can鈥檛 be predicted. But once an accident has occurred, the ensuing gridlock follows predictable patterns. 鈥淭he edge of a traffic jam propagates like a shock wave in a fluid,鈥 says Alexandre Bayen at CCIT.
Information on the density and speed of traffic is provided by existing 鈥渋nductive loop鈥 sensors built into the roadway. When a vehicle drives over the loop, the steel in its body causes an electric current running through the loop to change momentarily. The data gathered this way can then be fed into the Traffic Prediction Tool, a program developed by IBM.
The software also draws on GPS data from participants鈥 smartphones to learn their preferred travel times and routes, and uses this to provide predictions and recommend alternative routes tailor-made for that individual鈥檚 journey. For now, the California initiative provides predictions for only about a dozen participants. IBM hopes to expand the service to commuters worldwide.
聯Using GPS data from your smartphone, the software learns your preferred travel times and routes聰
Too many users in one area may be problematic, however. 鈥淚f everyone has the same information, then there may be an overreaction,鈥 says Moshe Ben-Akiva, a professor of civil and environmental engineering at the Massachusetts Institute of Technology. 鈥淵ou inform them of congestion in one location and the congestion just shifts to another route.鈥
IBM is now developing traffic software that takes this into account, by calculating how the predictions it has already made may cause congestion to occur elsewhere.
