Uber Reportedly Disabled Key Safety Feature Prior to Deadly Crash
The company may have been working on proprietary safety measures which required it to disable third-party software.
The fatal crash that claimed the life of a 49-year-old woman when an Uber self-driving vehicle collided with the pedestrian in Tempe, Arizona may have been preventable according to the company responsible for one of the key technologies developed for the autonomous Volvo.
According to a report by Bloomberg, the decision to disable the software before the crash may have ultimately resulted in the conditions which caused the pedestrian to be struck and killed.
Zach Peterson, a spokesperson for Aptiv PLC, spoke up to clarify that the collision-avoidance technology designed by the company for the particular Vovlo being driven was not faulty. “We don’t want people to be confused or think it was a failure of the technology that we supply for Volvo, because that’s not the case,” said Peterson in a statement to Bloomberg, “The Volvo XC90’s standard advanced driver-assistance system ‘has nothing to do’ with the Uber test vehicle’s autonomous driving system.”
Aptiv PLC obtains the sensors and other viable chips for its autonomous product from Intel's recent acquisition of Mobileye. It supplied the crash footage to the supplier who stated that the software “was able to detect Herzberg one second before impact in its internal tests,” despite the video being of poor quality. Though this may not be a substantial amount of time, the detection should have immediately triggered evasive actions to avoid the collision.
Given that the XC90 was traveling at 40 miles per hour, in the second that it was able to detect the pedestrian, the vehicle traveled less than 60 feet. In a test by MotorTrend on a 2016 XC90, stopping distance from 60 miles per hour was around 113 feet, meaning that although the vehicle may have not come to a complete stop, it may have impacted the pedestrian at a significantly lower speed, possibly preventing the death from occurring.
“The video released by the police seems to demonstrate that even the most basic building block of an autonomous vehicle system, the ability to detect and classify objects, is a challenging task,” wrote Amnon Shashua, Mobileye's CEO, who had also cited experience as a pillar of autonomous driving.
“It is this same technology that is required, before tackling even tougher challenges, as a foundational element of fully autonomous vehicles of the future,” he said.
Though Aptiv PLC isn't downplaying the severity of the accident, it wants to ensure that the public understands that the technology it develops, including collision-avoidance, lane keeping, and other autonomous driver aids, was not at fault in the accident. It is not clear if Uber's standard operating procedure dictates that the company disable third-party technology during testing, however it would make sense that if the company was developing a proprietary software, it would perform such actions. Uber declined to provide comment to Bloomberg on its inquiries regarding the accident.
One second to perform evasive maneuvers isn't a lot of time. Even if the vehicle was able to detect the driver and take appropriate action in the same period that Mobileye produced in its lab, the system's primary action would take would likely be to preserve the safety of the passengers aboard. LiDAR supplier Velodyne also confirmed that its sensors should have clearly imaged the pedestrian while she was crossing the street to allow software to take preventative actions. But somewhere, whatever the disconnect may have been, safeguards failed not only the vehicle but the woman who lost her life in the accident.
