A first-of-its kind study from Virginia Tech Transportation Institute found that the crash rate for self-driving cars is lower than the national crash rate.
The study was commissioned by Google, which is developing self-driving cars and testing them on roadways in Mountain View, California, and Austin, Texas.
The data took into account the severity of crashes, and it adjusted for unreported incidents. The researchers concluded that the national crash rate of 4.2 accidents per million miles is higher than the crash rate for self-driving cars, which is 3.2 crashes per million miles. The institute describes these findings as "reversing an initial assumption" that autonomous cars would have a higher incident rate.
The study also found some evidence that the self-driving cars were less likely to incur severe crashes, but it lacked sufficient data to draw a firm conclusion.
"These kinds of studies will be key," said Jeremiah Owyang, industry analyst with the firm Crowd Companies. Owyang, who spoke to Fast Company from the Detroit Auto Show, expects that research studies like these will bolster consumer confidence in self-driving cars.
He expects that companies in the space, like Google and Tesla, will leverage this data in their early advertising campaigns. Google is currently working to test and improve its cars' sensors in extreme weather conditions. In its latest report, Google stressed that its cars were equipped to understand rain.
Owyang said he'd like to see additional research into the psychological impact of these cars, and how that impacts crash rates. "How are people adjusting to seeing these vehicles on the road?"