In 2008, 7000 people died in car crashes, and of those deaths, over 700 at crashes in which a driver ran a red light. Researchers at MIT want to shrink those numbers with an algorithm to predict when a driver is going to run a light. Algorithms to do this already exist, but they do not have an accuracy of 85%.
Before we can get this in our garages though, vehicle-to-vehicle (V2V) systems will have to be implemented in cars. This technology is being explored by the Department of Transportation and many car manufacturers such as Ford. A V2V system allows cars to talk to each other so the vehicles and drivers can know better the distance between them, their speeds, and their directions. Benefits from such a system include gas mileage, by minimizing the need for minor speed adjustments, and safety with systems like MIT’s.