Experts See Autonomous-Car Revolution as Evolution

BERLIN –The first steps toward automated driving now are clear in Europe and North America.

But questions remain, some as simple as when drivers can safely watch a movie behind the wheel, others as complex as when the car will be a robot that needs no human operator to drive safely from point to point.

At a conference here sponsored by we.CONECT, an international facilitator of business-to-business networking, engineers from automakers, Tier 1 suppliers and technology companies present the pieces they are working on that one day will lead to autonomous driving where the car takes over everything.

“It is going to happen,” says Bakhtiar Litkouhi, a leader in autonomous-driving R&D for General Motors, “and electrification is going to happen.”

The auto industry is working on increasingly sophisticated advanced driver-assistance systems. Whereas the pioneering technologies, such as lane keeping, mainly offer safety benefits, the new systems also feature more comfort for drivers.

Traffic-jam assistance in which cars steer, brake and accelerate up to 18 mph (30 km/h) now is arriving on the market at Volvo and Mercedes-Benz.

Cars like that demonstrated by Audi at January’s Consumer Electronics Show that can find a parking place in a garage on their own are expected in several years, as are cars with so-called supercruise, which combine adaptive cruise control and lane-keeping on freeways.

While European automakers are leading the way in seeking solutions to the urban annoyances of parking and traffic jams, Cadillac could be the first U.S. company with hands-off supercruise giving motorists a break on long drives.

“What we have done so far is longitudinal control,” with adaptive cruise control and accident avoidance front and rear, Litkouhi says.

“Supercruise with a full speed range combining adaptive cruise control and lane centering is coming next. As soon as you provide steering, that is the first phase of automated driving.”

Automakers generally prefer the term “automated driving” rather than “autonomous” to distance public expectations from the idea that robot cars are just around the corner.

“Google pushed the idea” of autonomy with publicity over its self-driving cars that have logged millions of kilometers of tests with a driver monitoring the route behind the wheel, says Anders Eugensson, director-governmental affairs for Volvo Cars.

Bryant Walker Smith, a researcher at Stanford University’s Center for Automotive Research and the Center for the Internet and Society at Stanford’s law school, says consumers’ expectations of autonomy likely will increase because of publicity over the blind driver of a Google car, or a robotic-taxi demonstration in Berlin.

However, Smith cautions: “The view that people are getting is not realistic in the near term, (and) the view that it will answer 95% of all crashes sets up a difficult legal situation. If the cars don’t live up to the hype, if they crash, if they are not as cool, you ask, ‘Was it a bad design?’”

And in the courts, he says, “expectations shape what is considered a bad design.”

Some people involved in autonomy research believe a revolution is not far away.

Tim Kentley-Klay, founder of Australian start-up Zoox, says he is developing an autonomous electric vehicle that would operate in cities at lower speeds. Contacted and controlled by a smartphone app, it would provide personal transportation to people who would not need to own a car. He hopes to have a prototype made in 2016 and a vehicle on the market four or five years later.

However, a straw poll of attendees at the Berlin conference shows more than 75% expect an evolutionary path toward autonomy.

Decision-Making to Shift From Driver to Designer

The path starts with what NHTSA calls Level 1 autonomy, such as cruise control, and Level 2, such as supercruise, in which the car handles steering as well as longitudinal control. Both levels require drivers to be in charge all the time, deciding when to use or not use the feature.

At Level 3, cars will take responsibility for driving part of the time, such as in a platoon of vehicles connected by vehicle-to-vehicle communication. Level 4, the completely autonomous vehicle that does not need a driver, would arrive on shared streets afterward.

“We are gradually handing over control to the vehicle, and within the next 10 years we will make much more progress,” Stanford researcher Sven Beiker says.

Business consultancies Frost & Sullivan and Mira Consulting have released results of a study in which they expect to see the introduction of platooning, automated highway systems, emergency steering assistance and autonomous valet parking and retrieval by the end of the decade.

Both platooning and valet parking clearly are Level 3 features, which if implemented will add to automakers’ responsibility and liability.

“Decisions will shift from a human driver to a human designer in an office who was trying to anticipate the situation a vehicle might encounter,” says Stanford’s Smith. “Liability may be an obstacle, but the largest obstacle is uncertainty about the extent of that liability.

“Manufacturers are likely to take a greater share of crash damages. Optimistically, the total magnitude of accidents goes down considerably, so what that likely means is that manufacturers will be liable for a greater share of a smaller pool of damages.”

In the U.S., an activity is legal unless a law forbids it, which makes moving toward Level 3 technology easier. European nations are parties to the Vienna Convention on Road Traffic, which establishes standard traffic rules for countries that have signed the international agreement.

The convention says “every driver shall at all times be able to control his vehicle or to guide his animals,” says Volvo’s Eugensson, suggesting the language is a bit vague and may allow for interpretations of what constitutes “control.”

“Does this mean the driver is capable of being in control or is in control? If you say he has to be in control, there is trouble with ADAS (advanced driver-assistance systems) now,” he says, adding revisions to the convention are under review.

Meanwhile, European automakers are proceeding with developing the technology that would permit Level 3 automated driving – once they decide customers want it.

Audi has been a leader in demonstrating automated valet parking. At the January Consumer Electronics Show, the automaker demonstrated a car controlled by a smartphone parking itself in a narrow garage space, and it is developing a full valet service that lets an Audi drive into a garage and park itself.

Harald Avinger, a function developer at Audi Electronics Venture, says he thinks such a service could be legal in Europe because a parking garage is private property and not subject to terms of the Vienna Convention.

Platooning has been under development for years in Europe and has been successfully demonstrated in cars and trucks. V2V communication is critical for platoons of vehicles, which travel closely together and must respond to a sudden stop by braking at the same time. In a platoon, only the driver of the lead vehicle is in control, while occupants of all the other vehicles have no responsibility for safety.

BMW is among the automakers working on V2V communication. But, notes Michael Heimrath, manager of BMW’s connected-drive research, “There is always a delay of five years from research to cars.”

Continental Automotive is working on ADAS development with BMW, and Karlheinz Haupt, Continental executive vice president-ADAS, says the supplier expects by 2015 automated parking, in which a driver could park a car with a smartphone app.

Haupt also displays a roadmap outlining a new highway function in 2016, which could accommodate supercruise, and new city and highway functions in 2021 that anticipate Level 3 functions such as valet parking and platooning.

Stanford’s Beiker presents a timeline with supercruise arriving in 2016, self-parking in 2018 and platooning in 2020, with question marks surrounding technologies farther in the future.

The idea of autonomy goes beyond cars, Beiker says, noting mining and agriculture have used driverless vehicles for years. He says the auto industry could learn from them, and from research into public-transportation systems using driverless vehicles.

Emilio Frazzoli, a Massachusetts Institute of Technology researcher, has been studying Singapore to determine how many automated vehicles would be required to offer the personal mobility that today is provided by 800,000 cars. If people were willing to wait 20 minutes at rush hour, he says, 300,000 autonomous vehicles called by smartphone could replace those 800,000 cars.

“We are not just talking about auto 2.0,” he says, “but personal mobility 2.0.”