The chief executive officer of Israel’s Mobileye, which was acquired by Intel Corp. last year for a whopping $15.3 billion, is calling on automakers, regulators and technology entrepreneurs to put their heads together to set out a safety framework for fully autonomous cars.
Prof. Amnon Shashua, who co-founded Mobileye — a maker of advanced vision and driver assistance systems for use in autonomous vehicles — said the tragic death of Elaine Herzberg, the 49-year-old pedestrian who was hit last week by a self-driving Uber car operating in autonomous mode in Arizona, should not be used as an excuse to “stifle” important work in the development of self-driving cars.
Arizona’s governor, Doug Ducey, sent a letter on Monday to the CEO of Uber, Dara Khosrowshahi, saying he was suspending Uber’s self-driving car tests indefinitely in the state following the fatal crash and “an unquestionable failure” to comply with public safety standards, Bloomberg reported on Tuesday. The state of Massachusetts also last week asked companies to halt testing of their autonomous cars for a few days.
The accident could have vast implications for the development of self-driving vehicles, which have been touted as potentially safer than those driven by humans.
The testing of autonomous vehicles has been going on for months in the US, as automakers and technology companies compete to be the first with fully automated technology.
“More incidents like the one last week could do further harm to already fragile consumer trust and spur reactive regulation that could stifle this important work,” Shashua wrote in a March 26 editorial on the Intel website.
“I firmly believe the time to have a meaningful discussion on a safety validation framework for fully autonomous vehicles is now,” he wrote. “We invite automakers, technology companies in the field, regulators and other interested parties to convene so we can solve these important issues together.”
Shashua said that society expects self-driving cars to be held “to a higher standard than human drivers.” However, sensors’ detection and classification of objects, and interpretation of the information, remains a challenging task, and no shortcuts can be taken in safety-critical areas, he said.
“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,” he wrote.
“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 data sets of tens of millions of miles, and going through challenging preproduction validation tests” on dozens of production Advanced Driver-Assistance Systems (ADAS) programs, cannot be skipped, he said. “Experience counts, particularly in safety-critical areas.”
Transparency is also needed to gain public trust, he said. The developers of self-driving technologies must show how they ensure pedestrian and passenger safety. And the systems used in vehicles should rely on independent sources of information: camera, radar and Light Detection and Ranging systems (LIDAR), the remote sensing methods used by developers of autonomous cars.
“Fusing them together is good for comfort of driving but is bad for safety,” Shashua wrote.