Follow
The corners of streets are always hiding the truth from people. They can surprise pedestrians with an unexpected encounter or drivers with unforeseen accidents. The same debility is transferred to self-driving cars as well. How can they offer comprehensive safety since they have blank spots? However, a genius AI project comes to the rescue.
The CornerCamera Can Intercept the Penumbra Hidden Objects Leave on Walls or Other Obstacles
MIT’s Computer Science and Artificial Intelligence Laboratory has just devised a new algorithm that can refine autonomous cars. Their new superpower will be the ability to see the obstacles around the corners. Based on their registered data, they will be able to reconfigure the route to the safety of passengers.
The new technology is called CSAIL ‘CornerCamera’ and employs sensors to record the penumbra of objects behind an opaque barrier. This level of advanced artificial intelligence can replicate human’s ability to assess what lurks outside their line of sight. People are paying attention to reflections and shadows that can hint at the presence of a foreign body.
The lead author of the AI project, Katherine Bouman, claimed that they could use the obstacles to their own advantage. Therefore, by enabling self-driving cars to read penumbra, the walls and other obstructions can turn into ‘naturally-occurring cameras.’
The new system can intercept data in real-time. Therefore, the machine can measure the trajectory and speed of the hidden object. This type of information can teach cars how to distinguish a harmless traffic sign from a careless jogger who is about to jump from behind a truck in front of the vehicle.
“Even though those objects aren’t actually visible to the camera, we can look at how their movements affect the penumbra to determine where they are.”
The MIT’s AI Project Has Its Weaknesses
The paper indicates that this type of technology can fit a simple smartphone camera. This tells a lot about its adaptable side that can work well with any autonomous car.
On the other hand, the new device is not without its flaws. For instance, it needs certain lighting conditions to record data. Therefore, it can’t offer satisfactory results at night time or cloudy days.
Image source: 1
