Uber’s self-driving vehicle tests come to a halt in Arizona.
Last week saw the first recorded pedestrian death cause by an autonomous car. Elaine Herzberg was struck by one of Uber’s self-driving vehicles as she wheeled her bicycle across the road in the city of Tempe, Arizona.
Following the fatal collision, CEO Dara Khosrowshahi expressed his condolences to Herzberg’s family and Uber said it would be halting the testing of its autonomous vehicles in Arizona, Pittsburgh, San Francisco and Toronto.
Uber is collaborating with police to investigate how the incident could have happened.
Yesterday (26 March), Associated Press reported that the governor of Arizona, Doug Ducey, suspended the company’s self-driving privileges.
In a letter addressed to Khosrowshahi, Ducey said that video footage of the incident that led to Herzberg’s death raised questions about Uber’s ability to test its technology in Arizona. He noted that public safety should be the most important thing for those who are involved with self-driving cars, adding: “The incident that took place on March 18 is an unquestionable failure to comply with this expectation.”
Public trust damaged
The decision and the incident itself is likely to have done some major damage to the public perception of autonomous cars and the safety of such vehicles. Experts who watched the footage said that the sensors on the autonomous SUV should have seen Herzberg pushing her bicycle and braked before impact was made, even though the incident occurred at night.
According to The New York Times, internal documents from Uber showed that the testing system was experiencing a number of difficulties, including requiring much more human intervention than other self-driving ventures from other companies as well as difficulty driving through construction zones.
Spokesperson for the governor, Daniel Scarpinato, said: “I think we have prudent, responsible regulation. I think [what] you saw with today’s action is that we are all about public safety and accountability.”
Senior vice-president at Intel, Prof Amnon Shashua, said that there was a need for a “meaningful discussion on a safety validation framework for fully autonomous vehicles”.
He said that patience was key in order to ensure autonomous vehicles are safe. “Recent developments in artificial intelligence like deep neural networks have led many to believe that it is now easy to develop a highly accurate object-detection system and that the decade-plus experience of incumbent computer vision experts should be discounted.
“This dynamic has led to many new entrants in the field. While these techniques are helpful, the legacy of identifying and closing hundreds of corner cases, annotating datasets of tens of millions of miles, and going through challenging preproduction validation tests on dozens of production ADAS programs cannot be skipped. Experience counts, particularly in safety-critical areas.”