NTSB Says Uber’s Self Driving Vehicle Had No Way Detect Jaywalking Pedestrians In Fatal Crash

The National Transportation Safety Board has said that Uber’s autonomous vehicles had software flaws that made it unable to detect jaywalking pedestrians when it struck and killed an Arizona woman last year.

SO WHAT HAPPENED? The software Uber was using at the time of the crash “did not include a consideration for jaywalking pedestrians.” Essentially, the pedestrian was detected 6 seconds before the crash but the software was unable to process that she was a pedestrian because there was no crosswalk in the location where she was walking.

WHAT HAS CHANGED SINCE THEN? Uber has of course updated this. The new system would have detected her at 289 feet away or 4.5 seconds before impact. Under the old software, Uber didn’t detect the pedestrian until 1.2 seconds before collision.

ANYTHING ELSE? The NTSB also released that Uber self-driving vehicles were involved in 37 crashes in the 18 months before the fatal accident. Of those 37 incidents, 33 involved non-Uber vehicles hitting an Uber self-driving car. Two of the incidents involved Uber vehicle hitting the other vehicle, and the remaining two incidents were pedestrians damaging the stopped Uber vehicle.