Nvidia Woos Automakers With Self-Driving Car Chip

For all the talk about Google’s self-driving cars, there has been relatively little talk about the technology that powers them. Now, Nvidia is rolling out some tech that promise to bring the world closer to a future where auto-piloted cars that can see and detect the world around them are a reality. At the Consumer Electronics Show, the company rolled out Nvidia Drive automotive computers based on the Tegra X1 mobile super chip. Nvidia is offering two car computers: Drive PX is for developing autopilot capabilities and Drive CX is for creating advanced digital cockpit systems. Drive, which is based on the same architecture used in modern supercomputers, makes possible computer vision, deep learning and advanced cockpit visualization.
Last month, reports surfaced indicating Google has been working on getting its Android mobile operating system built directly into a variety of vehicles. Such a system would allow drivers to use mobile apps without having to connect their cellphones to their cars. Designed to support easier in-car use of audio and messaging apps, Android Auto has been touted as a way to integrate mobile devices with cars in a safer and more streamlined way.
"Mobile supercomputing will be central to tomorrow's car," said Jen-Hsun Huang, CEO and co-founder of Nvidia. "With vast arrays of cameras and displays, cars of the future will see and increasingly understand their surroundings. Whether finding their way back to you from a parking spot or using situational awareness to keep out of harm's way, future cars will do many amazing, seemingly intelligent things. Advances in computer vision, deep learning and graphics have finally put this dream within reach.”
Could This Go Mainstream?
We caught up with Charles King, principal analyst at Pund-IT, to get his thoughts on the new innovation. He told us, in essence, Nvidia's new Tegra X1 and Drive technologies are designed to make supercomputing mobile.
“On the chip side, the Tegra X1 delivers all the capabilities of the company's latest Maxwell GPU architecture in a mobile form factor and the same power envelope as its earlier Tegra K1 chips only with a massive performance increase. Nvidia says that single Tegra X1 chips can deliver a Teraflop of performance with 10 watts of power, equaling the performance of ASKI Red, the first supercomputer to achieve one Tflop performance in 2000,” he said.
Nvidia’s move to target automobiles as the first market for Tegra X1 and its new Drive PX and CX solutions make perfect sense to King. That’s partly because cars are rapidly becoming the world's most sophisticated mobile computing platforms, and Nvidia's Drive offerings will likely be attractive to numerous automakers looking to enhance their products' driving and media features/capabilities, he explained.
“While luxury vehicles are likely to provide the initial use cases for Nvidia's new solutions, I expect that it won't be long before the company's products make their way into commercial and industrial vehicles, and military applications. But that's just the beginning,” King said. “The concept of having what amounts to a Nvidia supercomputer under the hood of your car is likely to excite the imaginations of developers and commercial instincts of numerous vendors.”
Audi Is All In
The Nvidia Drive PX and Drive CX platforms include a range of software application modules from Nvidia or third-party solution providers. The Drive PX auto-pilot development platform and Drive CX cockpit computer will be available in the second quarter of 2015.
The concept is already gaining momentum with automakers like Audi. Ricky Hudi, executive vice president of Electrical/Electronics Development at Audi AG, said he sees Drive PX as a “crucial tool for further research and development."
The realization of smart automobiles requires high-performance processing solutions that enable sophisticated sensor fusion and innovative machine learning, said Thilo Koslowski, vice president and distinguished analyst at Gartner Inc. “This will create a new class of self-aware and ultimately self-driving vehicles that can assess, sense, understand and react to the state of their surroundings and occupants," he said.