Alan Ohnsman
The spot at 640 North Mill Avenue in Tempe, Arizona, where pedestrian Elaine Herzberg was struck and killed by a self-driving Uber test vehicle on March 18, 2018.
As federal investigators try to figure out why a self-driving Uber SUV failed to detect and avoid hitting a pedestrian walking her bicycle across an empty stretch of darkened road in Arizona, the company that made the vehicle’s high-tech LiDAR vision sensor says it’s “baffled” as to what went wrong.
Velodyne, the top supplier of LiDAR for self-driving car programs, said the Volvo XC90 involved in the crash appears to have one of its units. The company hasn’t dispatched staff to Tempe, Arizona, where a team from the National Transportation Safety Board, or NTSB, is investigating the crash that killed Elaine Herzberg, 49, on March 18.
“We are as baffled as anyone else,” Marta Hall, the company’s president and chief business development officer, told Forbes late Thursday. “We are at the service of engineers and investigators at NTSB and NHTSA as the findings and facts of the case are discovered. We haven’t sent engineers there, but are certainly willing to if requested.”
While about 100 people die daily in U.S. traffic accidents, Herzberg’s death is particularly notable as she appears to be the first pedestrian killed by an autonomous vehicle. For now, it’s unknown whether the accident was the result of faulty sensors, a software glitch, problems with the vehicle’s computer or something else entirely. A video of the accident from cameras on the Uber test vehicle released by the Tempe Police Department this week shows what looks like a total system failure.
Based on the footage, the automated Uber did not appear to see Herzberg and didn’t alter its course to avoid striking her as she crossed in the middle of the block. The video also shows that the Uber safety driver at the wheel, whose job is to take control if the system isn’t working properly, doesn’t appear to be paying attention until just before the vehicle hits Herzberg.
LiDAR, an acronym for light, detection and ranging, is a technology that shoots out pulsed laser beams to create 3-D “point cloud” images of a vehicle’s surroundings hundreds of feet away. It sees people, objects, animals, trees and other vehicles in daylight or at night that cameras or human eyes might not detect. Similarly, Uber’s vehicles also use radar to detect hard objects, like Herzberg’s bicycle, at longer distances, as well as multiple cameras.
“Our LiDAR is capable of clearly imaging Elaine and her bicycle in this situation. However, our LiDAR doesn’t make the decision to put on the brakes or get out of her way,” said Hall, who is the wife of David Hall, Velodyne’s CEO, founder and inventor of its LiDAR units. “We don’t know what sensors were on the Uber car that evening, if they were working, or how they were being used.”
(For more on Velodyne, see “How A 34-Year-Old Audio Equipment Company Is Leading The Self-Driving Car Revolution“ from the September 5, 2017, issue of Forbes.)
Uber didn’t immediately respond to an email inquiry to confirm its vehicle was using a Velodyne LiDAR and whether it has a technical relationship with the closely held San Jose, California, company that’s built the sensors for more than a decade.
“We at Velodyne are very sad, and sorry about the accident which took a life,” Hall said. Company founder and CEO David Hall, “believes this accident was not caused by the LiDAR. The problem lies elsewhere.”