Laser headlamps may be gradually finding their way onto high-end cars, but the smart lights of tomorrow might allow drivers to see clearly through snow and rain to potential perils ahead. A team at Carnegie Mellon’s Robotics Institute is cooking up a smart headlamp system that rather than blanket the road with light, instead assesses what’s in front of the car in real-time. By fragmenting the beams into a million smaller segments, each individually controlled, the car could potentially leave a falling droplet of rain in darkness.
Rain and snow are common causes of glare during nighttime driving, with light from the headlamps reflected back and potentially dazzling the driver to more dangerous objects up ahead. While some car companies have experimented with using radar-style technology to augment low-light vision, they’ve never been able to do away with that glare altogether.
The Carnegie Mellon approach is different, relying on high-speed processing, object tracking, and precise control over the light. A camera mounted at the front of the car is linked to a spatial light modulator, shutting portions of the beam on and off according to what’s in the field of view.
As well as following falling snow, sleet, and rain, the same technology can be used to track oncoming drivers and selectively dim the brightness that might otherwise dazzle them.
A further development in progress will use GPS to pre-warn the system of upcoming turns in the road, so that the lights can automatically shape themselves to its contours.
According to the researchers, the computer is swift enough to track an object moving at 50 mph across the front of the vehicle, just five meters away from the bumper.
The next goal is to shrink the hardware down to the point where it could fit into a headlamp itself, since the current prototype is externally mounted onto the hood of a car.