How safety shapes driverless car technology

How safety shapes driverless car technology
As driverless technology develops, one company is focusing solely on near fully autonomous vehicles (AV), a move they believe will address public safety concerns. Photo: National League of Cities

A new state law that takes effect in June creates a state work group that assists the Washington State Transportation Commission on making annual recommendations to state lawmakers for public policy on the use of driverless or self-driving vehicles.

Meanwhile, self-driving tech companies such as Waymo intend to have driverless vehicles available for public rider service by the end of the year with level four technology, which means the car can operate without human control under certain conditions. The highest is level five, where the vehicle is fully autonomous under all conditions.

At an April 18 event in Seattle hosted by the Institute for Legal Reform (ILR) and the U.S. Chamber Technology Engagement Center (C_TEC), Waymo Senior Counsel David Tressier outlined how it created the technology to make the vehicles work and in what ways public concerns over safety have driven development.

“We are very excited at the prospect of bringing self-driving car technology to the public and improving road safety,” he said.

The way to do that is by “building the world’s most experienced driver” through a combination of public road testing and aerospace simulation. Formerly the Google self-driving car project, Waymo later split off to form its own separate company in 2016. Since 2009, its AV software has driven five million autonomous miles on public roads.

Unlike other autonomous vehicle (AV) companies, Waymo is only focused on level four technology, a decision made based on its experience with Google’s self-driving car tests that had employees operating them.

“You have someone texting, not paying attention to the road, fumbling around with cords,” he said. “It’s actually very easy for humans to start trusting the technology. I made a decision at that point that we would only pursue level four autonomy, because it’s the safest.”

He added that complications arise with creating level 2-3 self-driving cars where the driver frequently takes and yields control, which also compounds liability issues.

One barrier they hope to surmount is public anxiety. A 2017 Pew Research Center survey of 4,135 U.S. adults found that “although they expect certain positive outcomes from these developments, their attitudes more frequently reflect worry and concern over the implications of these technologies for society as a whole.”

That worry was perhaps demonstrated after a recent deadly accident in Arizona involving a self-driving Uber vehicle. Although the driver was found to be not at fault, Uber quickly pulled those vehicles from the roads. In January, U.S. Transportation Secretary Elaine Chao announced plans to release this summer a new set of guidelines for autonomous vehicles, in part to promote greater safety.

“The deployment of self-driving cars is going to depend on the public acceptance and public trust, so we do feel a responsibility to start educating the public about how it works,” Tressier said.

He also argued that part of public acceptance comes with understanding how the technology can eliminate the kind of human error that causes around 90 percent of car accidents in the U.S. In 2016, there were almost 40,000 vehicle accident fatalities.

There’s also costs that might be saved. In 2010 alone, motor vehicle crashes cost the U.S. $871 billion in economic loss and societal harm. That same year, there were 32,999 fatalities, 3.9 million non-fatal injuries and 24 million damaged vehicles.

“The status quo is not acceptable,” Tressier said. “As we think then about the future of torts and the…liability regime for this emerging technology, I think it needs to be with the consideration of the backdrop of the status quo. (It) shouldn’t be acceptable.”

“The prospect and the promise of self-driving cars when deployed…is to reduce these traffic fatalities and increase road safety,” he added. “They don’t get drowsy…they can see 360 degrees, they can respond, and they can see up to three football fields in every direction.”

A 2017 RAND Corporation study concluded that waiting for “nearly perfect” driverless cars could waste an opportunity to reduce accident fatalities. “At best, fatalities are comparable, but, at worst, waiting has high human costs. Under none of the conditions we explored does waiting for significant safety gains result in fewer fatalities.”

Another RAND study released that year recommended “an approach in which AVs are introduced gradually as the vehicles meet a set of incremental, performance-based benchmarks” and “that the target benchmark of AV performance can determine the cap on vehicles’ deployment or, conversely, the number of vehicles desired can determine what benchmark should be set.”

Also released in 2017 was Waymo’s safety report, the first of its kind, which described in detail how the vehicles operate.

One feature of these vehicles is overlapping sensors, which Tressier said “will be important when it comes to liability, because it is so critical for safety. It (AV software) has to make sense of what it’s seeing in the world.”

4 COMMENTS

LEAVE A REPLY

Please enter your comment!
Please enter your name here