A lot of research has been going around sensors to be used in the self-driving cars that would enable them to have complete knowledge of their environment and surroundings. Several automakers and technology companies have come up with different solutions, some revolve around usage of Lidar, some rely on Camera and radar sensors but no such complete solution has come up that can solve the problem completely. In this scenario, some researchers from MIT claim to have developed a new kind of camera system called CornerCameras, that can detect objects or people in a hidden scene.
This system, developed by MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) can find application in wide range of products but for self-driving cars, it looks very promising and complementary to the products available now. The system uses information about light reflections to detect objects or people in a hidden scene and measure their speed and trajectory — all in real-time. Using video of the penumbra— the shadow which falls on the ground when light falls on the object— the system can stitch together a series of one-dimensional images that reveal information about the objects around the corner.