Driverless cars, robots can see around corners using radio waves

Researchers at the University of Pennsylvania have developed HoloRadar, a novel system that enables robots and autonomous vehicles to "see" around corners by using radio waves processed with artificial intelligence. Unlike traditional non-line-of-sight (NLOS) perception methods that rely on visible light, HoloRadar leverages the longer wavelengths of radio signals, which can reflect predictably off flat surfaces like walls, floors, and ceilings. This allows the system to reconstruct three-dimensional scenes beyond a robot’s direct line of sight, even in darkness or variable lighting conditions, thereby enhancing safety in cluttered indoor environments such as warehouses and factories. HoloRadar complements existing sensors like LiDAR rather than replacing them, providing an additional layer of perception that reveals hidden objects or pedestrians and gives autonomous systems more time to react to potential hazards. Tested successfully on mobile robots navigating hallways and building corners, the system was able to detect walls, corridors, and humans outside the robot’s direct view. Future research aims to adapt HoloRadar