Trends

Uber petitions Pennsylvania to resume self-driving tests

Uber said Friday it has asked Pennsylvania for permission to resume self-driving car testing on public roads and has improved the autonomous vehicle software, more than seven months after the company suspended testing after a deadly crash in Arizona.
The company disclosed in a report to the U.S. National Highway Traffic Safety Administration that it will resume testing with two employees in the front seat, will enable an automatic braking system at all times and more strictly monitor safety employees.
Uber has been testing its self-driving cars in manual mode on public streets.
In June, police in Tempe, Arizona, said a backup driver behind the wheel of a self-driving Uber was distracted and streaming a television show on her phone right up until about the time the car struck and killed a pedestrian walking across a street, deeming the crash that rocked the nascent industry “entirely avoidable.”
The crash was the first-ever death attributed to a self-driving vehicle and was seen as a significant setback for the industry, which is racing to get vehicles into commercial use.
Uber Chief Executive Dara Khosrowshahi said in a statement Friday the company “will go back on the road only when we’ve implemented improved processes.”
Uber said it now has real-time third-party monitoring of backup safety drivers, sets limits on time drivers can work per day and has improved training.
In July, Pennsylvania’s Department of Transportation said it was issuing new guidelines asking companies to submit details about testing but said it expected companies would comply.
The state said after it approves submissions it will send companies “an authorization letter.” Uber said it will not resume testing until it receives that letter.
Alphabet’s Waymo unit plans to launch a commercial ride-hailing service in Arizona this year, while General Motors is on track to roll out a similar service next year with vehicles without steering wheels or brake pedals.
The Uber crash raised significant questions about the performance of Uber’s software.
Uber said a key recommendation of an internal review after the Tempe crash was to improve the self-driving vehicles’ “overall software system design.” Uber said in a report released Friday the vehicle had improved “system latency. We are now able to detect objects and actors sooner and execute safe reactions faster.”
The National Transportation Safety Board said in May the Uber registered observations of the pedestrian about six seconds before impact, but the system did not determine that emergency braking was needed until 1.3 seconds before impact.
The NTSB said the vehicle had registered the pedestrian who was walking a bike across the road first as an unknown object, then a vehicle and lastly a bicycle.
Uber said it has a new approach to dealing with “handling uncertainty within the self-driving system.” Uber also has a new separate systems engineering testing team and plans a self-driving safety advisory board of outside experts. Uber said if a vehicle is uncertain about something in its view the software is now better positioned “to reason over many possible outcomes to ultimately come to a safe response.”
The NTSB also said that Uber had disabled a manufacturer-installed automatic emergency braking system in the 2017 Volvo XC90 while the car was under computer control in order to “reduce the potential for erratic vehicle behavior.”
Uber said Friday that braking system is now active whenever it is testing on public roads. It also filed a voluntary safety assessment with the National Highway Traffic Safety Administration on its self-driving efforts and made public a report from an outside law firm that reviewed the safety culture at Uber.
Source: VentureBeat
To Read Our Daily News Updates, Please Visit Inventiva Or Subscribe Our Newsletter & Push.

Related Articles

Leave a Reply

Your email address will not be published. Required fields are marked *

Back to top button

Adblock Detected

Please consider supporting us by disabling your ad blocker