The self-driving Audi turned smoothly onto El Camino Real, shifted into the middle lane and after a few stoplights found itself idling behind the distinctive plastic bubble of a Google autonomous car.
That sort of thing happens in Mountain View, the epicentre of autonomous driving tests. But whereas Google has not publicly outlined a path to commercialisation for its self-driving technology, the gear inside this particular Audi could hit the market in just a few years.
It was designed by Delphi Automotive, a British firm that is one of the world's largest auto parts suppliers, with help from Israeli vision-processing company Mobileye. Together, the two suppliers say they will have a self-driving platform ready for production in 2019, though automakers would likely need at least another year to incorporate it into their vehicles.
"We have tremendous respect for Google," said Glen De Vos, Delphi's vice president of services, from the Audi's backseat. "What they really demonstrated was that this technology is feasible. But we're in the business of taking technology and bringing it to market."
With the race to develop self-driving cars now at an all-out sprint, Delphi and Mobileye believe they possess an edge.
They have developed a system for crowdsourcing the hyper-detailed 3D maps upon which autonomous vehicles rely. Millions of non-autonomous cars that use Mobileye cameras for lane keeping or collision prevention will create a constant stream of data to map roads and potential obstacles, even temporary ones such as road repair crews or double-parked cars.
"High-definition mapping really begs for crowdsourcing," said Dan Galves, Mobileye's senior vice president of communications.
The platform, which will use Intel chips, will also be able to think through the situations each autonomous car will face on the road. Capable of performing 20 trillion mathematical operations per second, the system will try to anticipate what nearby cars and pedestrians will do and plan accordingly.
"You can't develop autonomous cars that just follow all the rules, because they'll just clog cities," Galves said. "The point is really providing the intelligence and the rules of breaking the rules, if you will – providing some human intuition into the vehicles."
Delphi and Mobileye plan to show off their integrated system at the Consumer Electronics Show in January, conducting demonstration drives on a six-mile course mixing freeways and city streets.
Delphi's tricked-out Audi handled both with ease during a 20-minute spin through Mountain View and Palo Alto on Thursday.
After leaving Delphi's Mountain View lab, the car navigated its way along busy streets, then onto Highway 101, then back to the lab without any input from the human behind the wheel. A screen on the dash showed a ghostly map of the road – blue lines against a black background – while cars and people picked up by the car's sensors appeared as Xs or clusters of dots.
When making right turns at lights, the car edged forward, only to pause for pedestrians in the crosswalk, including one very slow man carrying a bunch of bananas. Merging onto the freeway, the car accelerated slowly but surely, while avoiding the late-morning speed freaks blasting past.
"All of these things are programmable, which gives you some nice options," De Vos said. "If a driver wants more aggressive performance, that's programmable."
Unlike Google's cars, many of them stationed at a facility less than a mile from Delphi's lab, Delphi's self-driving vehicles hide their technology. There's no spinning mirror on top to give the car away.
While a trio of cameras sits just inside the windshield, the front radar lurks behind a plastic license plate. The front lidar – a system that operates much like radar, but with laser light instead of radio waves – is concealed behind a thin, black plastic panel right below Audi's trademark rings on the grille.
"When we started our autonomous driving program, we said, 'OK, we don't want it to look like a science experiment,'" De Vos said.
Delphi expects that autonomous technology will first be deployed in taxis – the company is already testing self-driving taxis in Singapore – as well as delivery vehicles. In both cases, the extra cost of self-driving technology, which could add US$50,000 (RM220,922) to the price of a vehicle, would be offset by the savings of not needing a human driver.
"When you take a US$100,000 (RM441,845) to US$150,000 (RM662,767) driver out, you can afford to pay for a lot of additional technology," De Vos said. — San Francisco Chronicle/Tribune News Service