Autonomous Cars Will Challenge Technology Infrastructure, Panelists Say

Jeff Klei, president of Continental AG's North American region, predicted that 54 million autonomous vehicles will be on the road by 2035.

But he cautioned that self-driving cars will need an accurate road map, plus reliable data from the infrastructure, such as stop lights, speed limits and lane closings, and also live traffic updates from other vehicles.

“You can’t just rely on vehicle sensors,” Klei said on the sidelines of the Automotive News World Congress. “You need data from the cloud to tell you what’s going on.”

Self-driving cars need all that data to figure out what hazards, if any, should be avoided on the road ahead beyond the range of a vehicle’s sensors. A couple of years ago, Continental successfully road-tested a static version of the map -- minus live updates from other vehicles -- with truckers in Europe.

Now it’s looking for customers that will install the map in their passenger vehicles. “It’s simple to do, once you get access to the data,” Klei said. “But we don’t have a time line” for production.

Dynamic maps

It may be awhile before dynamic maps are practical. To gain access to the data, automakers will need the cooperation of municipalities, which may or may not have a centralized traffic control system. Automakers also will need the approval of motorists, who will need assurance that their vehicle’s data won’t be misused.

Panelist Jim Sayer, a researcher at the University of Michigan Transportation Research Institute, said he believes motorists will share their data, up to a point. “We are giving up a lot of information for convenience, but not if our welfare is at risk,” he said.

Specifically, Sayer fears hackers could penetrate a city’s electronic network of signs and stoplights to gain illicit access to a vehicle’s data. “You could have a completely secure vehicle, but the traffic control systems are the weak link,” he said.

Sharing data

Once automakers gain the cooperation of municipalities and motorists, it would make sense to share all that data, noted panelist Danny Shapiro, senior automotive director at chipmaker Nvidia. If cars are hitting potholes, the road crews could fill them. And if cars are encountering black ice on the road, the city could dispatch salt trucks.

Likewise automakers might learn to share Big Data with each other, Shapiro added. “Each automaker has its own software team and it’s a race,” Shapiro said. “But you might use computer simulations to develop standard data bases, or regulators might create a basic performance test” for autonomous vehicles.

Given the cost of r&d, automakers may learn that they’ll have to cooperate as much as compete to make driverless cars a reality.