October 24, 2018 10:47 CET
JERUSALEM — Mobileye’s driverless car is part of Intel’s attempt to grab a stake in autonomous technology and the Israeli company, which the chipmaker bought for about $15 billion in 2017, is now reaping the benefits of being part of a larger company including having access to its engineering capabilities.
The car itself has developed a bit of a stubborn streak.
During a test drive in October, with a Bloomberg reporter sitting in the passenger seat, the car merged into traffic on a highway. Then, when drivers were reluctant to allow it to change lanes — a situation quite common on Israel’s roads — it made the move, smoothly but assertively.
Over the 20-minute drive, the driver did not touch the steering wheel or the brake once.
Like its rivals, Mobileye has experienced problems. At a May press event in the company’s hometown of Jerusalem, its camera-only vehicle ran a red light about a quarter of a mile from the company’s garage. The problem was blamed on the electromagnetic interference caused by wireless transmitters on cameras being used by a television crew. The company said it has since fixed the problem.
Mobileye has been keen to emphasize its focus on safety, without which society is unlikely to adapt to the idea of a car without a driver.
“When you help an industry accelerate the transition to an advanced data center platform, like assistive or autonomous driving, nothing is more strategic,” said Bob Swan, Intel’s chief financial officer and interim CEO.
Billions of dollars
Billions of dollars are pouring into autonomous technologies as automakers and technology giants battle to be the first to bring driverless cars to the masses. Indeed, Mobileye’s technology will also be used by automakers such as BMW and Fiat Chrysler Automobiles as the industry makes the move toward fully self-driving cars.
Despite two fatal autonomous car crashes involving vehicles run by Uber and Tesla, most contenders expect their cars to navigate cities at casual speeds within three years.
Without a safety model that persuades the public to accept autonomous cars, the whole industry is in danger of becoming ‘a science project,’ said Mobileye’s CEO Amnon Shashua, who created a safety code that sets clear rules for faults, based on a mathematical model.
The Israeli unit is also an experiment for Intel, as it tries to break its dependence on personal computer processors. Instead of absorbing its engineers and products into its own efforts, the world’s second-largest chipmaker left Mobileye intact. Underlining that independence, Mobileye’s most advanced chips are manufactured by competitors and not made in Intel plants.
In the second quarter Mobileye hit record revenue with a 37 percent, year-over-year increase, Intel said in July.
Six months ago, Mobileye repurposed some of Intel’s photonics engineers and tapped Intel engineers previously working on micro-radars for smart phones and put them on creating digital radars for a future generation of Mobileye chips.
‘There is lots and lots of potential for building the next generation sensor way beyond the current science of lidars,” said Shashua.
Unlike many of its autonomous competitors — such as Alphabet’s Waymo — Mobileye’s current autonomous prototype does not use lasers or radars. They will be added once development is completed, which is expected by the end of the year. Meanwhile, Mobileye’s main business is supplying automakers with a ‘driver assist’ system — a camera and custom chip that allows it to predict when a driver is about to hit something.
Via Mobileye, Intel has the highest number of systems on the road, albeit at lower levels of autonomy as they pursue an incremental approach, unlike companies such as Nvidia, which is aiming to jump straight to fully autonomous systems.
In a move that could help it win automotive market share, Mobileye’s customers will be able to add their own code to its next chip. Mobileye’s range of EyeQ chips are currently a closed system, meaning the client purchases the chip and software as a single package.
Allowing automakers to write their own code — a process known as open architecture — could let items such as lidar processing or driving software be added on top of Mobileye’s existing technology.
By making its chips customizable, Mobileye is attempting to demystify its technology to its growing base of major automobile clients, who are pursuing a range of methods in the race to launch an autonomous car.
The change is likely to be significant and grow the Mobileye market, said Michael Ramsey, research director for Automotive and Smart Mobility at Gartner Inc. “One of the big complaints has been that Mobileye is a black box, the car companies don’t know how it works and can’t make adjustments,” said Ramsey. “Companies’ biggest frustration with Mobileye was because of that.”
The ability to open up Mobileye’s technology came after Intel’s acquisition was completed in August 2017, when 200 of the multinational’s engineers moved over to the Israeli company and joined its Eyeq5 chip development team, said Shashua.
“The way Mobileye works is as a closed system,” Shashua said. “But given those 200 engineers from Haifa we created another market in which we can offer open compute to customers who want to write their own code.”
Even before Mobileye’s newest chip, still under development, hits the market, the company has been announcing deals. In May, it signed a contract to supply a European automaker’s eight million cars with self-driving technologies. Shashua declined to elaborate and added that the company “has been closing many deals.”
Contact Automotive News