There's nothing more irritating than driving out of a traffic jam and not seeing a cause of it. Where was the crash? Where was the person fixing a flat tire? Imagine if all that could be solved with a simple cruise control programming.
MIT professor Berthold Horn used his own experiences driving the notoriously backed-up I-93 in Massachusetts. In cars currently equipped with adaptive cruise control systems, there's a radar watching the speed and distance of the car in front. Car in front slows down, you slow down and then the car behind you also slows down.
Horn's idea is to go one step further and use a system to monitor the speed of the cars both in front and behind you, then forcing you to drive a speed somewhere in between of those other two cars. That means the cycle of flashing brake lights that makes people panic and slow down would end.
Apparently, his algorithm works even when factoring driver reaction times and variable driving speeds.
And all of this can be accomplished with the sorts of cameras already being used by automakers to detect pedestrians, rather than using more expensive radars.
It would be interesting to see this in action, though, and determine how much driver input there really is. The bigger issue, which a UC San Diego professor says, is getting people convinced they need systems that monitor all of the drivers around them, not just themselves.
You can see Horn's entire paper, Suppressing Traffic Flow Instabilities, here.