A new kind of radar could allow unmanned vehicles to navigate safely in bad weather. Electrical engineers have developed a clever way to improve the imaging capabilities of existing radar sensors to accurately predict the shape and size of objects being observed.
Developers from the University of California, San Diego said the new system has already performed well when tested at night and in foggy conditions. The engineers will present their work at the Sensys conference from November 16-19 in Japan.
Severe weather conditions pose challenges for self-driving cars. These vehicles use technologies such as LiDAR and radar to “see” and navigate, but each has its own drawbacks. LiDAR, which works by reflecting laser beams off surrounding objects, can draw high-resolution 3D images on a clear day, but it does not work well in fog, dust, rain or snow. On the other hand, a radar that emits radio waves can see in any weather but captures only a partial image of the road scene.
New technology from UC San Diego engineers dramatically improves radar visibility.
“It’s a LiDAR-like radar,” explains Dinesh Bharadia, professor of electrical and computer engineering at the University of California, San Diego. He noted that this is an inexpensive way to get better performance in self-driving cars during bad weather. “With our technology, there is no need to use expensive LiDARs.”
The system consists of two radar sensors located on the bonnet and spaced at a distance of the average vehicle width (1.5 meters). The location of the two radar sensors is key – they allow the system to see more space and detail than a single radar sensor.
During test drives on clear days and nights, the system, like the LiDAR sensor, perfectly determined the dimensions of vehicles moving in traffic. Later, engineers “hid” another vehicle with a fogger, and their system accurately predicted its three-dimensional geometry. LiDAR sensor failed test.