Between 2014 and last November, Google’s self-driving cars had 272 failures and would have crashed 13 times if a human had not intervened, according to a new document.
In a report submitted to the Department of Motor Vehicles (DMV) in California, all companies with driverless cars on the state’s roads had to submit accident details concerning when vehicles suffered a “disengagement”, or an incident when a human had to step in and correct the car.
Over a 14 month period, 49 self-driving cars were used by Google, including the “koala” cars and a fleet of modified Lexus RX450s.
The cars drove 424,000 miles by themselves by had 341 incidents in which the car handed control back to the driver or they had to step in to prevent an accident.
The 272 failures mentioned involved things like communication failures, or faulty sensors readings, for example. In those cases, the car would provide an indication to the driver that they should take over.
Google said its drivers had to take over the car many thousand times however, and that it was only reporting on selected incidents as the company interprets California’s rules omit times when the car could have coped on its own without a driver.
However John M. Simpson, Consumer Watchdog’s Privacy Project director said. “The DMV got it exactly right and is putting our safety first. How can Google propose a car with no steering wheel, brakes or driver when its own tests show that over 15 months the robot technology failed and handed control to the driver 272 times and a test driver felt compelled to intervene 69 times?”
Six other car companies (VW/Audi, Mercedes Benz, Delphi Automotive, Tesla Motors, Bosch and Nissan) are required to submit reports by January 1, 2016. Until now only Google has made the results public.
SOURCE: Consumer Watchdog.
Larry Banks is a keen follower of technology and finance. He has worked for a variety of online publications, writing about a diverse range of topics including mobile networks, patents, and Internet video delivery technologies.