Autonomous cars: Uber reveals ‘valuable lessons’ in safety report
uber self-driving

Autonomous cars: Uber reveals ‘valuable lessons’ in safety report

Uber Advanced Technologies Group has released a report that outlines the company’s commitment to it’s self-driving vehicle strategy and what it’s doing to insure the safe development of autonomous cars.

Titled ‘A Principled Approach To Safety’, the voluntary safety self-assessment was developed in line with the National Highway Traffic Safety Administration’s guidance.

The 70-page document is intended to speak to multiple audiences, including the public, fellow road-users and potential users of self-driving technology, policymakers (including legislators), regulators, local officials and other self-driving vehicle developers.

Uber believes that competitive pressures have made sharing information on progress in development challenging. Yet transparency into developments and progress are important to earn and increase public confidence in this technology and, in turn, its ability to deliver on the potential benefits.

Learning from the past

The report follows a fatal crash in March, in which an Uber autonomous test car struck and killed a homeless woman, 49-year-old Elaine Herzberg.

Commenting in the report, Uber CEO Dara Khosrowshahi addressed the accident, and outlined the motivation behind the document:

“We are deeply regretful for the crash in Tempe, Arizona, this March. In the hours following, we grounded our self-driving fleets in every city they were operating. In the months since, we have undertaken a top-to-bottom review of ATG’s safety approaches, system development, and culture.

“We continue to support the National Transportation Safety Board’s investigation into the Tempe crash. We have taken a measured, phased approach to returning to on-road testing, starting first with manual driving in Pittsburgh.”

We committed to deliver this safety report before returning to on-road testing in self-driving mode, and will go back on the road only when we’ve implemented improved processes.

Uber’s self-driving safety principles

The U.S. Department of Transportation and its National Highway Traffic Safety Administration (NHTSA) identify 12 safety elements for safe self-driving. Uber leans heavily on these in its safety principles, which inform its development program:

  • Proficient: In the absence of hardware faults, how do we demonstrate that our system performs more safely than human drivers, using credible and tractable performance metrics?
  • Fail-Safe: How do we ensure that the system responds to a malfunction that could result in harm to a person by transitioning to a state which reduces the risk of harm? Under what circumstances will we allow a risk to persist?
  • Continuously Improving: How do our development processes capture, consider, and respond to undesirable or unexpected system behaviour?
  • Resilient: How do we prevent, protect, and/or warn against potential harm that arises when our technology is used counter to its design or purpose by external actors?
  • Trustworthy: How do we create and maintain two-way dialogue with our riders, regulators, legislators, other road users, and advocacy organizations and provide them with evidence of safe performance?

Internet of Business says

Uber is hoping its recent advances in object and risk processing will see it rewarded with permission to role out further on-road testing.

Critically, it is working on what it calls “next generation ultrasonic sensors (USSs)”, which uses echolocation to range objects, and will provide near-ranging sensing of people and objects within five metres of the vehicle.

The technology can assist with stopping and starting, lane changing, and parallel parking. These sensors will be distributed across the front and rear fascia and the starboard and port side sills.

More than 3 million driver and delivery partners on the Uber network enable approximately 15 million trips every day. While the company emphasises that the transition to driverless vehicles will be a long road, and that its drivers are an integral part of the business, the economic benefits to be had from not needing to pay for drivers are potentially massive.

The likely end goal will make past friction between Uber and its drivers, and with local governments, pale by comparison, though it will be no less than the wider cultural and regulatory changes that will be demanded by ubiquitous autonomous vehicles.

The report follows a recent MIT survey into the surprisingly complex ethical questions raised by self-driving vehicles.