Waymo has traveled 7.1 million driverless miles. How does his driving behavior compare to that of humans?

Waymo has claimed for years that its self-driving vehicles have the potential to be safer than humans. Now the company says it has the data to back it up.

Waymo analyzed 7.13 million fully self-driving miles in three cities – Phoenix, Los Angeles and San Francisco – and compared the data to human driving benchmarks to determine whether its cars were involved in fewer injury-causing and police-reported crashes. And it was the first time the company studied miles of fully self-driving operations, rather than a combination of autonomous and human-controlled driving.

Waymo analyzed 7.1 million fully self-driving kilometers in three cities

The conclusion? Waymo’s self-driving cars were 6.7 times less likely than human drivers to be involved in a crash resulting in injury, or an 85 percent reduction compared to the human benchmark, and 2.3 times less likely to be involved in a police-reported accident, or a reduction of 57 percent. That translates to an estimated 17 fewer injuries and 20 fewer police-reported crashes compared to if a human driver had driven the same distance in the cities where Waymo operates.

Waymo’s analysis comes at a fraught time for autonomous vehicles. The company’s main competitor, Cruise, has halted operations across the country after an accident in San Francisco resulted in a pedestrian being dragged twenty feet by one of the company’s self-driving cars. Cruise allegedly withheld video footage of the incident from regulators and now faces up to $1.5 million in fines from the state. Lawmakers and union activists are calling for a crackdown on autonomous vehicles, regardless of company ownership.

But Waymo said the timing of the release of its security analysis – which comes at a time when its main rival is facing its worst crisis in years – is merely a coincidence. Still, Waymo’s director of safety research and best practices, Trent Victor, said the data can help the public understand that not every company developing autonomous vehicles is the same.

“What we would like to do is give a clearer picture so people can see the difference” between Waymo and other AV companies, Victor said in a briefing with reporters. “Another difference is that we are scaling up responsibly.”

The company uses public data to form its conclusions and invites other outside researchers to replicate the results. Waymo and other AV operators are required to report any accidents to the National Highway Traffic Safety Administration as part of the agency’s efforts to shed light on self-driving cars in the US.

Waymo’s millions of miles were not entirely without incident. The company said that overall, over more than 4.5 million miles in all three cities, its vehicles were involved in only three crashes that resulted in injuries: two in Phoenix and one in San Francisco. According to Kristofer Kusano, a safety researcher at Waymo and co-author of the study, all three injuries were minor.

Waymo’s analysis comes at a fraught time for autonomous vehicles

Yet that is considerably less than the crash risk for human drivers. Another way to look at it is to look at the accident rate per million kilometers driven. The human benchmark is 2.78 incidents per million miles. Waymo’s benchmark for its self-driving vehicles was just 0.41.

One of the biggest challenges for Waymo was controlling several factors when comparing its vehicles to human drivers. To make a fair comparison, Waymo had to address statistical biases in the data, such as human drivers not reporting minor accidents, or differences in driving conditions.

For example, Waymo’s vehicles operate in geofenced areas in the three cities they serve, excluding highways. Human drivers do not avoid these types of roads. Human drivers also tend not to report certain low-level accidents, such as minor fender benders. Waymo, on the other hand, is required by law to report any contact event with another vehicle, no matter how minor. As a result, Waymo had to adjust its model to take these factors into account.

Waymo’s analysis included “underreporting of adjustments for crashes reported by police or those derived from naturalistic driving study databases,” the company said, citing “a literature review of twelve previous studies and one book comparing AV and human crash rates .”

Those millions of kilometers were not entirely without incident

“The goal was to say what the human crash population would have looked like if it had been driving under similar conditions as the automated vehicle,” said John Scanlon, a safety researcher at Waymo and co-author of the study.

The analysis follows a study Waymo published with Swiss Re that found the company’s self-driving vehicles reduced the frequency of personal injury claims by 100 percent, compared to Swiss Re’s human baseline of 1.11 claims per million miles.

The current state of the self-driving car industry is very unsettled. AVs are active in a small handful of cities, but many people say they are skeptical of the technology and won’t necessarily trust it over their own driving habits. The negative attention surrounding the Cruise incident, as well as the recent recall of Tesla’s Autopilot, are just the latest in a series of bad headlines that have led many to conclude that self-driving cars are a fad, or worse, more dangerous than human drivers.

Waymo has gone further than other AV companies in using data and statistical analysis to argue for the safety of autonomous vehicles. Waymo emphasizes that self-driving cars are necessary as an antidote to the crisis of traffic fatalities, of which there are approximately 40,000 in the US every year. The company often points out that driverless cars never get drunk, tired or distracted and can avoid the human errors that so often lead to accidents and deaths.

The current state of the self-driving car industry is very unsettled

Last year, the company produced two scientific papers comparing the performance of autonomous vehicles to human driving. The first analyzed reaction times when an accident is imminent, while the other presented a new methodology to evaluate how well autonomous driving systems avoid accidents. Waymo has also tried to measure the safety of its AVs by simulating dozens of real-world fatal accidents that have occurred in Arizona over nearly a decade. The Google spin-off found that replacing both vehicles in a two-car collision with robotic vehicles would eliminate almost all deaths.

It is an indisputable fact that there are far fewer AVs on the road than human-driven vehicles and therefore less data from which to draw conclusions. People drive nearly 100 million miles between fatal accidents. Some experts argue that we will need hundreds of millions of miles of autonomous vehicles before we can make more meaningful comparisons about safety.

But Waymo’s efforts appear to be a step in the right direction. “These reports represent a good faith effort by Waymo to evaluate its security [autonomous driving system] comparable to the safety of human driving,” said David Zuby, principal investigator of the Insurance Institute for Highway Safety. “The results are encouraging and represent a step in our evolving understanding of ADS safety.”

Other researchers agree. “This provides the strongest evidence I’ve seen yet that ADS have a lower crash rate than humans,” said Carol Flannagan, a researcher at the University of Michigan’s Transportation Research Institute. “I think the more complicated question is what this means in the broader process of determining when ADS is ‘safer than humans.’”

Leave a Reply

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