Hot on HuffPost Tech:

See More Stories
AOL Tech

IBM's Traffic Jam Tech Predicts Gridlock Before It Happens

traffic jamReal-time traffic updates delivered to your GPS or cell phone are not particularly new. But, as Scientific American points out, there is a major limitation to these systems. Namely, most drivers are already on the road, and often already stuck in traffic, when they receive these alerts. This is because they're little more than live traffic reports and cannot predict what traffic may look like in the future. IBM has developed a system, which has been tested in Singapore, Finland and New Jersey, that is able to predict traffic conditions up to an hour ahead of time.

The new technique uses predictive modeling, based on a number of factors, to accurately predict traffic volume and speed. By combining live data from sensors and cameras, GPS transponders in taxis, analysis of historical traffic information, roadwork conditions and weather forecasts, IBM was able to accurately predict conditions over 85-percent of the time in the three trial markets. Alerts of traffic problems are then broadcast to both navigation systems and road signs, at which point they can even give accurate estimates of when traffic will return to normal.

IBM is still working to optimize the system in order to avoid simply shifting congestion from one road to another. It is also in talks with cellular providers to integrate measures of cell phone density on roads for more fine-grained predictions. IBM has signed contracts to deploy the system with two U.S. transportation authorities, but the specific locales have not yet been announced.

Of course, predictive modeling technology has uses far beyond traffic management. Washington, D.C. is using the IBM tech for flood and sewer management, and the company is also testing a program that could predict the effectiveness of various HIV treatments. [From: Scientific American]

Tags: cartech, GPS, gridlock, ibm, predictive modeling, PredictiveModeling, top, traffic, traffic management, traffic prediction, TrafficManagement, TrafficPrediction, transportation