Princeton University is developing a new radar technology that will enable self-driving car AI systems to read unseen objects around corners. Experts have found a way to create this low-cost radar, which, if passed testing, could help improve safety on the road, especially for pedestrians and motorcyclists. The radar’s function is said to be similar to looking at blind spots around corners through a reflective glass usually placed at intersections. Researchers at the university designed the technology using Doppler radar, which has been developed for a decade, making it easily integrable into modern vehicles.
According to RideApart, Doppler radar is commonly used in warfare and weather forecasting needs. However, radar systems are now being used as vehicle aids. Doppler radar reflects radio waves from surfaces such as buildings and parked cars. The reflection is not vertical, but triangular. The emitted signal will hit hidden objects in curved areas. The signal is then reflected back to the detector inside the car, allowing the system to see objects around corners and determine whether the object is moving or stationary. “This will allow cars to see obstacles that cannot be detected by lidar and cameras. It’s useful for self-driving cars passing through intersections,” said Felix Heide.
Radar sensors are relatively cheap, especially compared to sensors, and can be mass-produced. Researchers describe the radar system as capable of distinguishing objects including cars, bicyclists, and pedestrians.
Fangyin Wei, a computer science student at the university, said that this new system can distinguish between bicyclists and pedestrians. Wei said the system can be adjusted to detect vehicles. First, he needs to determine if there is anything there. If there is something, is it important? Is it a bicyclist or a pedestrian? Meanwhile, Heide explained that researchers intend to conduct additional research to develop radar as a means of improving road safety.
