With car manufacturers pushing autonomous cars as the future, it’s easy to be swept up in the hype. But just how safe are they and when will they surpass humans?
Not for a while if a report by the California Department of Motor Vehicles is anything to go by. As required by state law, companies involved in testing autonomous vehicles have to publish the good and bad – and it’s not all good.
Google, for instance, has detected failures with its autonomous technologies 272 times between September 2014 and November 2015, which lead to the driver having to take control of the vehicle, 13 of which would have otherwise ended in a crash. Drivers responded to the immediate takeover in 0.84 seconds.
Another 69 instances involving Google’s 53-car fleet saw the driver feel like it was necessary to intervene.
Google admits autonomous cars have a long way to go, but said in its report that disengagements are ‘a critical part of the testing process to expand the software’s capibilities and identify areas of improvement’. In other words, eggs have to be broken to make an omelette.
Let’s also not forget its fleet travels 30,000-40,000 miles a month in self-driving mode – considerably more than the average driver covers in a year.
Seven of the 11 companies approved to test autonomous cars in the state of California were required to disclose disengagements by the 1st of January 2016.
A total of 2,894 disengagements were reported by all companies, 1,051 of which came from Mercedes-Benz and its two vehicles, 625 from Bosch and its two vehicles, 260 from VW and its two vehicles, 106 from Nissan and 511 from Delphi Automotive.
Google self-driving car program director Chris Urmson said in a blog post: “Thanks to all this testing, we can develop measurable confidence in our abilities in various environments.
“This stands in contrast to the hazy variability we accept in experienced human drivers – never mind the 16-year-olds we send onto the streets to learn amidst the rest of us.
“Although we’re not quite ready to declare that we’re safer than average human drivers on public roads, we’re happy to be making steady progress toward the day we can start inviting members of the public to use our cars.”
Four of Google’s self-driving cars were involved in four accidents between September 2014 and May 2015, it was reported.
There are plenty of issues on the path to fully-autonomous cars becoming mainstream beyond the technology itself. For starters, liability is an issue – who is responsible for a crash, the manufacturer or those onboard? How will insurers deal with it?
Then there are the surrounding laws. Should at least one passenger of an autonomous car have a driving licence, just in case something goes wrong (and inevitably it will)? Should you be able to get a lift home if you’re under the influence of alcohol?
Critics have argued it will be difficult for humans to relinquish control of their vehicles, especially if autonomous cars are only marginally safer. That’s assuming the technologies are affordable, although a few manufacturers have promised affordability.
So how soon will autonomous cars arrive? Renault-Nissan boss Carlos Ghosn believes it will have a car on the market in 2020, contrasting strongly with Tesla boss Elon Musk’s claim of 24-36 months for one of its cars to drive itself across the US.
Responding to questions at the 2016 Detroit Motor Show, Ghosn said: “What is an autonomous car? If it’s a question of being autonomous on one lane on a highway or maybe changing lanes then yes this [is something we can do in] 2016 or 2017.
“But if you’re talking about autonomous driving in a city, with crossroads or the car making decisions in complicated situations then frankly I don’t think it’s going to be ready before 2020.”
A recent study found the majority of motorists want a steering wheel in a self-driving car. Be sure to check out our guide to self-driving cars for more information.
Leave a Reply