When a self-driving car boots up and takes in the world around it, its vision is nothing like what a human driver finds familiar. Instead, a self-driving car uses data collection via radar, cameras, and lidar to "see" the world around it.
How can humans begin to understand what the machine sees? Make the human a self-driving car, of course.
That's exactly what Moove Lab and Meso Digital Interiors did with a new project. The two companies created a cart of sorts, but it features all the equipment a self-driving car on the road today utilizes. With only the data collection as "sight," the test required human drivers to actually pilot the cart.
The human test driver wears an Oculus Rift virtual reality headset that displays the data and information gathered from the self-driving car's sensors. With only the data, the human drives. The main display comes from a 3-D-depth camera to map the environment in real time. While this happens, lidar sensors calculate the distance between the vehicle and other objects with light. From the video, we see percentages calculate the likelihood of what's ahead, which also helps the driver make his or her decision.
It helps answer a question of "how do self-driving cars see the world?" And if you want our answer, we'll use the standard method instead of the enormous amounts of numbers in front of our eyes.