Self-driving cars have been proclaimed as the solution to car accidents, removing the human risk factor from the equation, but those working on the technology say it won’t be perfect.
“People shouldn’t think that there will never be a crash,” said Ron Medford, director of safety on Google’s self-driving car project. “You’re going to be much, much better than a human, but you’re not going to be absolutely perfect.”
Medford’s comments came up during an autonomous vehicle discussion panel which took place at the Connected Car Expo ahead of the LA Auto Show press days last week. Jeff Klei, North America president of Continental AG, also took part in the panel.
Google has been testing the self-driving technology for at least three years and the only reported accident since testing occurred while the car was under manual driver control.
Medford also explained that Google believes that vehicles won’t have to rely on internet connectivity in order to self drive.
“Google does not believe that we need to have connectivity in order to have full autonomy,” he said. “We can do it without it.”
Klei expressed similar concerns about the success rate of autonomous vehicles. “Just like any environment, there will be accidents,” he said. He also stated that partially autonomous technology should start rolling out to vehicles as soon as 2016, but fully self driving cars are a bit further down the road, some time around 2020.
One such future technology was showcased in Ford’s new Edge Concept, with an active steering system. Mated to current auto-brake functionality it would allow the car to not only apply the brakes by itself, but actually operate the angle of the wheels to steer the car around an object in its path.
While experts agree that self driving cars will cut down on the 1,100 injuries and 10 fatalities that occur each day due to distracted driving, it’s clear that expectations have to be set realistically about the new technology.