Yes, we have prototyped cars that can drive by themselves, but without smart vision, they cannot really tell the difference between a crumpled paper bag on the road, which can be run over, and a rock that size, which should be avoided.
The idea that a robot will become more aware of its environment, that telling it to 'go to the kitchen' means something - navigation and understanding of the environment is a robot problem. Those are the technological frontiers of the robotics industry.
We're doing a lot of work on self-driving cars. We do not currently have cars in the U.S., but we plan to, for development and testing. I think we are within striking distance of making self-driving cars a reality, and these would be powered by deep learning.
The most important thing for building a robot that you can interact with socially is its visual attention system. Because what it pays attention to is what it's seeing and interacting with, and what you're understanding what it's doing.
All the technology going into self-driving cars is robotic technology. It's not automotive. That explains why some of the traditional automotive players didn't develop this technology.
Well, you can say there is a self driving car. I'm seeing the automation of vehicles. Really, computer-assisted driving. I think that is really interesting to us because we are taking all of the sensors technologies and putting them in cars and making people safer.
A self-driving car will have 200-plus CPUs. That's a data center on wheels.
To make a vehicle autonomous, you need to gather massive streams of data from loads of sensors and cameras and process that data on the fly so that the car can 'see' what's around it.
When cars have the sensory systems around them, GPS intelligence, they're looking at the world not only in visual spectrum, but infrared, ultraviolet and everything else that's going on and they've got reaction times in microseconds. Not a tenth of a second. They're a hundred thousand times faster.
