You’ll be able to buy a car that can drive itself under most conditions, with an option for override by a human driver, in 2020, according to the median estimate in a survey of 217 attendees of the 2014 Automated Vehicles Symposium. By 2030, the group estimated, you’ll be able to buy a car that is so fully automated it won’t even have the option for a human driver.
Though 2020 is just six years away, there remains a lot of debate over how the industry is going to get there. Most auto manufacturers are incrementalists, adding automated features such as adaptive cruise control, self-parking, and traffic-jam assist, two or three at a time. Google and some others in Silicon Valley, however, are more interested in producing highly or even fully automated cars as soon as possible.
The Society of Automotive Engineers and National Highway Traffic Safety Administration have slightly different definitions of different levels of automated cars. But both basically agree that a “partially automated car” can take over some driving functions such as speed and steering but can’t actually drive itself; “highly automated” can drive itself under most conditions but has a human override, and “fully automated” can drive itself without a human override.
Don Norman, a human-factors engineer from UC San Diego, seemingly endorsed Google’s strategy in a keynote speech Wednesday that argued that even highly automated vehicles might be too dangerous for people to use. “I strongly favor full automation,” he said, but feared that highly automated vehicles might find themselves unable to handle some condition and give the human drive an inadequate amount of time to safely take over.
Airplanes have been highly automated for years, Norman pointed out, but if a plane that is 30,000 feet up suddenly decides it can’t handle current conditions and demands that a human takes over, the human pilot has several minutes before the plane might crash. An auto driver in the same position might have only a fraction of a second.
As I indicated here July 16, another source of debate is whether cars should come with vehicle-to-vehicle (V2) communications. This would allow, among other things, cars to operate in “platoons” of several to many cars each. Since the cars would be in constant contact with each other, they could close the gaps between them and greatly increase highway capacities.
Though 2020 is just six years away, there remains a lot of debate over how the industry is going to get there. Most auto manufacturers are incrementalists, adding automated features such as adaptive cruise control, self-parking, and traffic-jam assist, two or three at a time. Google and some others in Silicon Valley, however, are more interested in producing highly or even fully automated cars as soon as possible.The Society of Automotive Engineers and National Highway Traffic Safety Administration have slightly different definitions of different levels of automated cars. But both basically agree that a “partially automated car” can take over some driving functions such as speed and steering but can’t actually drive itself; “highly automated” can drive itself under most conditions but has a human override, and “fully automated” can drive itself without a human override.
Don Norman, a human-factors engineer from UC San Diego, seemingly endorsed Google’s strategy in a keynote speech Wednesday that argued that even highly automated vehicles might be too dangerous for people to use. “I strongly favor full automation,” he said, but feared that highly automated vehicles might find themselves unable to handle some condition and give the human drive an inadequate amount of time to safely take over.
Airplanes have been highly automated for years, Norman pointed out, but if a plane that is 30,000 feet up suddenly decides it can’t handle current conditions and demands that a human takes over, the human pilot has several minutes before the plane might crash. An auto driver in the same position might have only a fraction of a second.
As I indicated here July 16, another source of debate is whether cars should come with vehicle-to-vehicle (V2) communications. This would allow, among other things, cars to operate in “platoons” of several to many cars each. Since the cars would be in constant contact with each other, they could close the gaps between them and greatly increase highway capacities.
by Randal O'Toole, Kurzweil | Read more:
Image: Harbrick
