REAL LIFE NEWS: UBER’S SELF-DRIVING FATALITY COULD HAVE BEEN CAUSED BY SOFTWARE BUG
by Hazed
The tragic incident where an Uber self-driving car killed a pedestrian back in March (see http://www.ibgames.net/fednews/2018/180325/official04.html) may have been caused by a bug in the software.
According to anonymous sources, the car’s sensors did detect Elaine Herzberg as she crossed the road pushing her bike, but the software classified the detection as a “false positive” and decided not to stop for her.
Software designers need to carry out a delicate balancing act between allowing too many false positives and too many false negatives. If the software is programmed to be too cautious, it will stop too many times for no good reason, making the ride slow and jerky. But if the software is tuned in the other direction it will risk ignoring a real object.
Hence the tragic encounter between car and pedestrian.
The reporter who broke the story said, “There’s a reason Uber would tune its system to be less cautious about objects around the car. It is trying to develop a self-driving car that is comfortable to ride in.”
He added, “Uber had been racing to meet an end-of-year internal goal of allowing customers in the Phoenix area to ride in Uber’s autonomous Volvo vehicles with no safety driver sitting behind the wheel.”
Uber hasn’t commented on this report.
