Anyone who has had to drive a car during a heavy downpour or snowstorm knows how difficult it can be to see the road in front of you. Instead of improving visibility, the light reflects off of the snow or rain rather than the scene in front of you, making it more difficult to see rather than less. According to www.phys.org, a team headed by Professor Srinivasa Narasimhan from Carnegie Mellon University in Pittsburgh is working on improving headlights so that they are capable of avoiding the falling snow and rain, thus dramatically improving visibility and potentially saving lives.
These “smart headlights” are actually a system that includes a DLP (digital light processing) projector similar to the projectors commonly used for presentations, a camera, and a beamsplitter. The beamsplitter allows the camera and projector to use the same “light channel,” making for a much more efficient design by reducing the workload on the image processor. A desktop PC was used as the image processing system. All components were COTS (commercial off the shelf).
The concept behind the system is simple: rather than directing light everywhere, the smart headlights track the precipitation long enough to determine a projected path, then directs the projector to project light where the precipitation is not, thus dramatically cutting down on glare. The system was able to pass through 90 percent of the light emitted from the projector in heavy rainfall conditions during computer simulations, though the simulated system was operating almost ten times faster than the physical system constructed by the researchers. It is important to keep in mind that the off-the-shelf nature of the laboratory setup is not nearly as fast or efficient as a custom-designed system would be, but even a small decrease in glare could yield significant benefits under actual driving conditions.
Despite the clear benefits of the “smart headlights” system, it is not equally capable in all situations; the system relies on its ability to predict the path of the precipitation, so turbulent or rapidly varying wind conditions, unpredictable car maneuvers, and precipitation consisting of small particles or droplets would all present difficulties. Still, the team believes that a custom designed system and better predictive algorithms will yield improvements in system performance. Despite that, don’t expect to drive a car equipped with these headlights any time soon; it will take a number of years before the system can be tested or employed in moving vehicles and even longer before it would be employed in commercially-available vehicles.