Nvidia has developed a supercomputer that will enable fully driverless vehicles.
The artificial intelligence supplier introduced its Drive PX Pegasus, the latest iteration of its automated driving platform, at its GPU Technology Conference in Munich on Tuesday. The platform can process 320 trillion operations a second, compared with 30 trillion operations on the preceding Drive PX 2 platform, which the company says is enough to power full self-driving, or Level 5 autonomy.
"After the first generation platform, we realized we needed a lot more performance," said Danny Shapiro, director of automotive at Nvidia, on a call with reporters. "This is a true supercomputer for the car."
Twenty of the 225 companies using Nvidia's Drive PX platform are developing driverless taxis, the company said. To safely eliminate a human driver, these cars require a suite of camera, lidar, radar and ultrasonic sensors that produce mass amounts of data -- as much as 1 terabyte per second.
Shapiro said Drive PX Pegasus can process this data instantaneously on a platform the size of a license plate, using less power than previously necessary to enable the car to make sense of its surroundings.
"This system is going to save thousands of watts off of what is in these cars today," he said.
Many of the companies developing on the Nvidia platform include automakers, suppliers and startups such as Zoox, Optimus Ride and NuTonomy.
"NuTonomy is building for Level 5 and Pegasus is the kind of platform that will be required to support these types of systems," said Karl Iagnemma, NuTonomy CEO, in a statement.
Drive PX Pegasus will be available in the second half of 2018, and companies developing autonomous vehicles on the Drive PX 2 platform will be able to upgrade to the new system once it is out, Shapiro said.