MIT says algorithm can predict red light runners
In a paper that will appear in the journal IEEE Transactions on Intelligent Transportation Systems, Jonathan How, professor of aeronautics and astronautics at MIT, and three of his colleagues say they have come up with an algorithm that can predict whether an oncoming car is about to run a red light one or two seconds before a possible collision.
That might not sound like a lot of time, but it could be enough to save a life.
The complex algorithm can quickly compute -- down to milliseconds -- the likelihood of a vehicle running a red light based on its rate of deceleration as it is approaching the intersection.
How and his team applied the algorithm to more than 15,000 vehicles at a busy intersection in Christianburg, Va., that was already outfitted with instruments that monitor vehicle speed and location as well as when the lights turned red. They found that they were able to correctly predict who would run a red light 85% of the time.
That accuracy is the highest that's been tallied so far.
The next step is getting the information about who is going to run a red light to other drivers on the road so that they know not to drive into that intersection.
According to How, that might be possible in "smart" cars of the future.
"If you had some type of heads-up display for the driver, it might be something where the algorithms are analyzing and saying, 'We're concerned,'" he said in a release put out by MIT. "Even though your light might be green, it may recommend you not go, because there are people behaving badly that you may not be aware of."
However, How said your car would only know that the driver of another car is about to run a red light if your cars were already "talking" to each other -- wirelessly trading information like speed and position data.
The technical term for this type of communication is vehicle-to-vehicle or V2V, and apparently it's on its way. Both the U.S. Department of Transportation and the Ford Motor Company are currently exploring V2V technologies.
Our question is will this technology, even realized in its fullest, be more effective than the doomed red-light camera pictures that L.A. was taking?
Only time will tell.
-- Deborah Netburn
Image: Could an algorithm have prevented this 1988 collision? Only time will tell. Credit: Insurance Institute for Highway Safety