Researchers at a number of US universities have created an imaging system designed to allow self driving cars to ‘see around corners’. The system uses conventional lidar and cameras. The researchers have developed AI that can interpret scattered light patterns that allow it to predict a car approaching from behind a corner.

Using lasers, the imaging system bounces light off visible walls and onto objects to create a scattered pattern. 

The image is then reconstructed by the AI system. ‘Noise’ patterns are removed, and images can be produced that are so accurate that even letters of 1 cm tall can be read.

According to researchers, the system can produce sub-millimetre detail of objects hidden behind a corner from a metre away. Combined with other imaging sensors, very large scale reconstructions can be created. 

Users of the technology include not only making it safer to drive through intersections, but also parking and driving away. 

The technology also has applications in medicine, particularly medical imaging, industrial robotics, and defence.

Universities involved in the research include Stanford, Rice and Princeton.

“Compared to other approaches, our non-line-of-sight imaging system provides uniquely high resolutions and imaging speeds,” said Christopher Metzler of Stanford University. “These attributes enable applications that wouldn’t otherwise be possible, such as reading the license plate of a hidden car as it is driving, or reading a badge worn by someone walking on the other side of a corner”.

The research has been funded by the Defense Advanced Research Projects Agency, DARPA.