MIT researchers have developed a self-driving car that can see around corners, thanks to a system of video cameras that focus on a very specific area. In this example, the system analyzes the ground where two perpendicular roads or paths meet, and then uses cameras to capture a series of video frames that can be intelligently processed and analyzed even when the vehicles are in motion. This enables them to detect the presence of shadows and changing lighting conditions from frame to frame. Read more for a video and additional information.
This data can also be used to predict the position and speed of an approaching object hidden around a corner, thus giving self-driving cars ample time to slow or completely stop if necessary. When compared to laser-based LiDAR hardware, this camera-based system was quicker to analyze by more than half a second. It will need a lot more testing until the software is ready for real roads.