Rand Study Calls for Greater Self-Driving Vehicles Safety Standards

Autonomous vehicle developers need to create a common set of safety metrics and terminology for self-driving cars, and this also requires cross-collaboration between industry partners and the government agencies that regulate them.

These were the recommendations of a Rand Corporation report funded by Uber’s autonomous vehicle division, which found that the focus of such industry efforts should be building trust with a public skeptical of AV technology.

Entitled “Measuring Automated Vehicle Safety,” the 91-page document, available free online, presents a framework for measuring safety in AVs that could be used by automakers, policymakers and the public.

In addition, the report shows how reduction of risk is possible through simulation, closed courses, and public roads with and without a safety driver, and at different stages of development and deployment.

“We are showing the art of the possible,” Rand Senior Policy Analyst Marjory Blumenthal told The Washington Post. “There are many instances in which it is possible to have a safety-relevant measure. What makes sense will vary with the circumstances.”

The report concluded information sharing should be encouraged and should precisely incorporate metrics such as measurements, context and data security, among other factors.

While some industry leaders in the nascent, yet highly competitive, self-driving vehicle market may balk at sharing what they consider trade secrets, the report seemed optimistic on this front.

“There is hope of more collective action among competitors,” the report noted, “what some might call cooperation.”

The Rand report also called for additional research into how self-driving vehicles communicate with their surrounding environment. This dovetails with a study released this week by the UK’s International Underwriting Association (IUA) indicating that lack of connected car infrastructure is a major barrier to wider adoption of AVs.

Still, the market is awash about what exactly autonomous vehicle technology can do, which is starting to confuse consumers.

Recent reports have shown consumers are increasingly confused as to what advanced driver assistance systems (ADAS) actually do, and automakers, in an effort to differentiate their feature sets, may be misleading drivers as to how capable these technologies actually are.

Earlier this month American automaker Ford called on autonomous vehicle manufacturers and developers to share ideas for what could become a standard way for driverless cars to communicate with other road users.

Ford also wants to bring the rest of the AV industry into the conversation over how self-driving cars should communicate what they’re doing. It invited developers to share ideas to create a global industry standard — one that might, or might not, look like Ford’s own solution.

Meanwhile, South Korean carmaker Hyundai is buying into AI technology that “understands” people’s intentions in a bid to accelerate its ambitions in autonomous driving.

The carmaker’s corporate and open innovation business, Hyundai Cradle, has announced it is investing in Perceptive Automata to develop AI software for cars and automated systems.

For its part, British automaker Jaguar Land Rover has enlisted the help of a team of cognitive psychologists to better understand how vehicle behavior affects human confidence.

The goal is to find out if it is beneficial to provide humans with information about a vehicle’s intentions, or whether simply letting a pedestrian know he or she has been recognized is enough to improve confidence.

— Nathan Eddy is a filmmaker and freelance journalist based in Berlin. Follow him on Twitter.