Ford wants to keep drivers alert on the long road to autonomous cars

21.10.2014
Self-driving vehicles will make the driver redundant, but long before that, smarter cars may leave the driver thinking about other things.

Ford is already studying that problem, anticipating an evolution toward autonomous cars that will take a lot longer than projects by the likes of Google may suggest. For now, the motorist is still in charge -- with some help.

"We still have a driver-centric model. We still think the driver needs to be engaged," said Don Butler, Ford's executive vice president of connected vehicle and services. Drivers and passengers are "more comfortable" when they have some control over the vehicle, he said during a session at GigaOm Structure Connect in San Francisco.

Ford thinks autonomous cars will be real, but for now it's adding automated assistance one feature at a time. For example, it makes vehicles that can park themselves, and that can keep drivers from drifting into the next lane.

There are technological, regulatory and social challenges to making fully autonomous cars a reality, Butler said in an interview.

Cars may not be ready to drive themselves in all areas and in all weather conditions, he said. Less structured environments outside of cities may be more of a challenge, for example. Some governments may not be ready to deal with the implications of regulating autonomous cars. And most consumers don't yet trust their cars to drive safely, Butler said. Putting the vehicle in charge full-time even raises ethical questions, such as what's the least harmful action when a collision is inevitable.

Instead of building fully autonomous cars, Ford plans to gradually add more technology to help drivers. Examples available now include automatic parallel parking and "lane assist," which detects if the car is leaving its lane and alerts the driver through sound or vibration. Like other carmakers, Ford also offers adaptive cruise control that prevents getting too close to the car ahead, and forward collision warning.

Future automated features may involve "crowdsourcing" data from many cars in the same area. For example, if one car's wheels slip in snow or ice on a stretch of road, it could share that information so other vehicles could automatically adjust for the conditions, Butler said.

But moving gradually to cars that drive themselves raises another issue: How to make drivers pay attention when they have less to do.

Ford's exploring that problem, Butler said. As features such as lane assist and auto-parking are added, the company is trying to make sure drivers understand their limitations. And Ford is looking at issues such as cognitive load, or the number of tasks a driver's brain can deal with at one time.

An early example of making sure drivers are paying attention is drowsiness detection, Butler said. Lane-assist is its own form of drowsiness detection, using the driving itself to determine whether a motorist is dozing off. The auto industry is developing other techniques, including in-car cameras that watch the driver's eyelids and seatbelts that sense breathing and heart rates.

Stephen Lawson