An Uber self-driving car has hit and killed a woman crossing the street in Arizona, marking the first time a self-driving car has killed a pedestrian and dealing a potential blow to technology which is expected to transform transportation.
- A self-driving Uber car has hit and killed a pedestrian in Phoenix
- It is unclear if the car, which was travelling at 65 kilometres an hour, slowed before the collision
- The Volvo was in fully autonomous mode with a driver behind the wheel
Uber said it was suspending North American tests of its self-driving vehicles, which have been going on for months in the Phoenix area, Pittsburgh, San Francisco and Toronto.
Elaine Herzberg, 49, was walking her bicycle outside the pedestrian crossing on a four-lane road in Tempe, Phoenix at about 10:00pm on Sunday (local time) when she was struck by the Uber vehicle traveling at about 65 kilometres per hour, police said.
The car was in autonomous mode with an operator behind the wheel and police were unsure whether it slowed down before the collision.
Ms Herzberg later died from her injuries in hospital, police said.
Local television footage of the scene showed a crumpled bike and a Volvo XC90 SUV with a damaged front.
Volvo confirmed its vehicle was involved in the crash but said the software controlling the SUV was not its own.
The US National Highway Traffic Safety Administration and National Transportation Safety Board said they were sending teams to investigate the crash.
Uber chief executive Dara Khosrowshahi expressed condolences on Twitter and said the company was working with local law enforcement on the investigation.
Self-driving cars are a billion-dollar business
Carmakers and technology companies are fiercely competing to be the first to release self-driving technology.
So-called robot cars, when fully developed by companies including Uber, Alphabet Inc and General Motors Co, are expected to drastically cut down on motor vehicle fatalities and create billion-dollar businesses.
But Sunday’s accident underscored the possible challenges ahead for the promising technology as the cars confront real-world situations involving real people.
US politicians have been debating legislation that would speed the introduction of self-driving cars.
“This tragic accident underscores why we need to be exceptionally cautious when testing and deploying autonomous vehicle technologies on public roads,” Democratic senator Edward Markey, a member of the transportation committee, said in a statement.
Last Friday, Uber and Waymo urged Congress to pass sweeping legislation to speed the introduction of self-driving cars.
Some congressional Democrats have blocked the legislation over safety concerns, and Monday’s fatality could hamper passage of the bill.
The US government has voluntary guidelines for companies that want to test autonomous vehicles, leaving much of the regulation up to states.
The US Department of Transportation is considering other voluntary guidelines that it says will help foster innovation.
Uber has said its ability to build autonomous cars is essential to its success in the rapidly changing transportation industry.
The company envisions a network of autonomous cars that would be summoned through the Uber app that would supplement — and eventually replace — human-driven cars.
Uber has logged 3.2 million self-driving kilometres through December.
The company has more than 100 autonomous cars testing on the roads of the greater Phoenix area, the company’s prime testing ground due to the state’s loose regulations and hospitable weather.
How does the technology work?
Self-driving cars are programmed to recognise road markings and signs, as well as GPS technology that conveys speed limits, road closures and traffic conditions.
Experts claim they are unlikely to make mistakes, but the technology falls down when human error by others is involved. This means if a human driver in another car is about to cause a crash, for example, it is unlikely a driverless car can do much to avoid a collision.
However the cars are programmed with “relative negative consequence of impact” technology, which is designed to choose the “least bad” option.
So if a dog runs onto the road for example, but swerving would mean hitting a child on the opposite footpath, the driverless car can distinguish between life forms and avoid hitting the child.
It can be very effective …
Numerous videos have been posted online showing driverless cars successfully avoiding collisions.
Supporters of the technology claim the cars are much safer than driver-operated vehicles as they eliminate human error, which is the main cause of accidents.
The cars can react much faster than human drivers to dodge potentially dangerous situations and have the ability to predict when a collision may occur.
… but accidents still happen
Self-driving cars being tested routinely get into fender-benders with other vehicles.
Last week, a self-driving Uber crashed with another vehicle in Pittsburgh, local news reported. There were no injuries.
A year ago, Uber temporarily grounded its self-driving cars for a few days following a crash with another car in Tempe.
The company has been the subject of a number of complaints about its autonomous vehicles, but it has said the cars were being driven by a human driver at the time of the incidents.
California is among the states that require manufacturers to report any incidents to the motor vehicle department during the testing phase.
As of early March, the agency received 59 such reports.
Concerns over the safety of autonomous vehicles flared in July 2016 after a man was killed while travelling in a Tesla partially self-driving car which collided with a truck in Florida.
Safety regulators later determined Tesla was not at fault.