Srinivasa Narasimhan and colleagues at Carnegie Mellon University's Robotics Institute invented a smart headlight system that improves visibility by redirecting light to shine between rain particles.
The system is made up of a camera that tracks the motion of raindrops and snowflakes falling in front of the car's headlights. The images are then sent to a processor, which uses a computer algorithm to predict where the particles will fall a few milliseconds later so it can redirect light rays to eliminate the glare from falling drops of rain.
"If you're driving in a thunderstorm, the smart headlights will make it seem like it's a drizzle," Narasimhan said in a statement.
The system is also capable of detecting oncoming cars and can direct headlight beams away from the eyes of drivers moving in the opposite direction.
Lab tests have demonstrated the efficacy of the system, but researchers say they need to work out a few problems before it can be used in cars.
"The system isn't perfect--in heavy rain accuracy is at 70 percent (that is, it removes 70 percent of the rain from view) at roughly 18 miles per hour. At 60 miles per hour, that drops to just 15 or 20 percent. But even 20 percent is a fairly good bump in visibility--certainly better than zero percent. The next step is to make the system better at accounting for car movements that aren't simply straight forward (presumably compensating for turning or lane changes and the like)."
Researchers are now engineering a smaller version of the smart headlight to be installed in a car for road testing
Photo via Carnegie Mellon