Intel to Team With Delphi and Mobileye for Self-Driving Cars

Today’s automobiles are often described as computers on wheels, for the scores of processors and chips they use to control everything, including the transmission, brakes, power windows and navigation system.

The advent of self-driving cars may require the equivalent of a supercomputer on wheels. Which is why three technology companies in the field — Intel, Delphi Automotive and Mobileye — plan to collaborate in an alliance to be announced Tuesday.

For Intel, especially, it will be an effort to catch up in autonomous vehicles, a field where some chip makers have made deeper inroads.

The processing power required to scan the road, identify pedestrians and fuse images from radar, cameras and other sensors — all in real time — is spurring a race to provide increasingly complex computer brains that will dwarf those found in cars today.

The competition is reflected in the partnership being announced on Tuesday, in which Intel will provide specialized computer chips to Delphi, an auto supplier, and Mobileye, an Israeli company that specializes in vision systems that have been used in some of the autonomous-driving systems made by Tesla Motors.

Within about two years, Delphi and Mobileye hope to offer automakers a system that can give less expensive cars and trucks the intelligence to drive themselves. At the center will be a package of Mobileye and Intel chips capable of computing about 20 trillion mathematical operations a second, Glen DeVos, Delphi’s vice president of engineering and services, said in an interview on Monday.

A later version of the system, he said, will aim to have two to three times that processing power.

“To be able to do all the computation you need for a fully automated vehicle, you can almost never have too much processing power,” he said.

An Intel spokeswoman confirmed the partnership. She said Delphi and Mobileye would begin using the Core i7 Intel chip, and later would use a more powerful and unnamed processor to be unveiled in a few weeks.

The partnership is the latest by Intel in its bid to muscle into the rapidly expanding automotive chip business. Nvidia, Qualcomm and a few other companies are ahead of Intel, which once dominated the personal computer business but has struggled to duplicate that success in other areas, including mobile devices and automobiles.

Intel announced this month that it would invest $250 million in start-ups working on automated-driving technologies. In July it formed a partnership with Mobileye and the German automaker BMW to provide chips for a self-driving car that BMW intends to begin producing by 2021.

“Intel is looking to get into the automotive space because the demand for processing power in cars is going to skyrocket,” said Michael Ramsey, an analyst at Gartner Group who follows automated driving trends.

According to Gartner, the automotive semiconductor business generated revenue of nearly $30 billion in 2015, up from nearly $15 billion in 2003.

Specialized chips are one of the essential technologies needed to make self-driving cars a reality. To perform safely and competently, autonomous vehicles need radar to detect obstacles, cameras to identify pedestrians and the color of traffic lights, highly detailed 3-D maps to determine the vehicle’s precise position and superfast computer brains to pull all of this information into split-second decisions.

“You’re taking in vast amounts of visual data, and you have to process it really fast, with no delay,” Mr. Ramsey said. “It all has to happen in real time.”

Intel faces formidable competition. Nvidia makes a processing unit that Audi is putting into its newest models, and another that Tesla has just started using its cars.

The Nvidia device used by Tesla, called the Drive PX 2, can compute 24 trillion operations a second. Nvidia recently demonstrated a more powerful version called Xavier.

“Intel is a very powerful company,” said Danny Shapiro, Nvidia’s senior directive of automotive technology. “But they are coming late to the game. We have been in this space a long time already.”

Mr. DeVos said Delphi chose to work with Intel because the chip maker had a plan to produce increasingly powerful automotive processors, and the scale to make the system affordable for mainstream cars.

“It’s all about getting to mass production,” he said.

Delphi is using its own radar technology and Mobileye’s image processing system, with self-driving algorithms developed by Ottomatika, a company spun off from Carnegie Mellon University.

Ottomatika, which Delphi acquired last year, is a source for many of the basic technologies used in autonomous vehicles.

Correction: November 29, 2016
An earlier version of this article misstated the number of operations per second that Nvidia’s Drive PX2 can compute. It is 24 trillion operations a second, not 24 million.

http://www.nytimes.com/2016/11/29/business/intel-to-team-with-delphi-and-mobileye-for-self-driving-cars.html?WT.mc_id=SmartBriefs-Newsletter&WT.mc_ev=click&ad-keywords=smartbriefsnl&_r=0