Cars that predict red-light runners coming soon?

All of us have felt that pang when the light turns green and we edge into a hectic intersection: Will someone run the red and smash into my car?

But what if you could prevent an accident by knowing when another driver might speed through a red light? Researchers at MIT say they've taken the first step in doing just that.

Jonathan How, the Richard Cockburn Maclaurin professor of aeronautics and astronautics at the Massachusetts Institute of Technology (MIT), and three other researchers  say they've developed an algorithm able to consistently predict a collision one or two seconds before it happens. "Smart cars of the future" theoretically could use such algorithms to help motorists anticipate and avoid these accidents, How says in a statement.

Get free custom rate quotes online, now.

"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,'" says How. "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."

Fewer accidents means potential car insurance savings

Beyond preventing injuries and saving lives, a reduction in crashes might also lower car insurance costs for insurers.  There would be fewer accident claims requiring payouts for repairs or medical expenses, says Pete Moraga, a spokesperson for the Insurance Information Network of California.

He adds that policyholders could benefit because auto owners insurance premiums tend to rise, sometimes significantly, after you've been in an accident.

"Keep in mind that insurers write policies based on what they see their risks as being now and in the future," Moraga says. "If there was less risk perceived by an insurer, that might result in a discount for the consumer."

Willem Rijksen, a spokesman for the American Insurance Association, says he hopes that MIT's research eventually proves valuable for both motorists and auto insurance companies.

"Auto and highway safety continue to be top priorities for the industry, as attention to safety measures can prevent needless losses, both financially and emotionally," he says. "While reducing accidents reduces insurance costs, increasing driver safety remains the primary concern."

Collisions at intersections affect millions of motorists annually. In 2008, there were 2.3 million intersection-related car crashes in the U.S., resulting in nearly 7,000 deaths, according to the National Highway Traffic Safety Administration (IIHS). More than 700 of those fatalities were due to drivers barreling through red lights.

The IIHS says half of the people killed in the accidents are not the red-light runners, but other motorists, passengers and pedestrians.

How says the algorithm could be used in evolving vehicle-to-vehicle (V2V) communications systems currently being studied by the Department of Transportation and some auto manufacturers, including Ford. How says cars using the algorithm could  "talk" to each other through sophisticated electronics with specialized wireless systems and collision-avoidance systems.

The MIT team refined its algorithm by studying more than 15,000 cars with speed and location monitors as they went in and out of a busy intersection in Christiansburg, Va. Researchers were able to correctly predict 85 percent of the time when a crash might occur.

How and his team will report their full findings in an upcoming edition of the journal IEEE Transactions on Intelligent Transportation Systems.