Autonomous vehicles have widely been speculated as becoming an industry changer in regards to safety, as driver mistakes account for almost all deadly crashes. However, according to a new study by the Insurance Institute for Highway Safety, only a third of all crashes could potentially be avoided if automated systems operate too similarly to human drivers.
Although autonomous vehicles will, at some point, be able to spot hazards and obstacles and react to safely avoid them much faster than humans, and they won’t operate with any distractions, stopping crashes altogether is still a huge challenge.
“We’re still going to see some issues even if autonomous vehicles might react more quickly than humans do,” said vice president of research for the institute and co-author of the study, Jessica Cicchino. “They’re not going to always be able to react instantaneously.”
Because of this, the Institute’s study found that although driver error is the ruling factor in 9 out of 10 crashes, only a third of those were due to mistakes that automated vehicles could avoid due to their superior perception abilities. To avoid the remaining two-thirds of crashes, automated vehicles would need programming to specifically prioritize safety over everything else–including speed.
“Building self-driving cars that drive as well as people do is a big challenge in itself,” said lead author of the study and research scientist for IIHS, Alexandra Mueller. “But, they’d actually need to be better than that to deliver on the promises we’ve all heard.”
The institution studied over 5,000 crashes caused by “sensing and perceiving” errors like driver distraction, failing to spot hazards in time, or impaired visibility, as found by the National Highway Traffic Safety Administration. Crashes were also distinguished by “incapacitation,” such as drivers being impaired by drugs or alcohol, medical problems, or severe fatigue. The study found that autonomous vehicles can prevent these issues.
“It’s likely that fully self-driving cars will eventually identify hazards better than people, but we found that this alone would not prevent the bulk of crashes,” Cicchino explained.
Some of these unavoidable challenges for current self-driving technology include planning errors, like driving too fast for certain road conditions; execution errors, like unsafe evasive maneuvers; or misjudging another vehicle’s speed.
Error-free, 360 degree perception by these vehicles is key, Cicchino said. She referenced one example: if a cyclist or other vehicle suddenly enters an autonomous vehicle’s path, the self-driving car could likely not stop or steer away fast enough.
“Autonomous vehicles need to not only perceive the world around them perfectly, they need to respond to what’s around them as well,” Cicchino said.
In the study, crashes as a result of sensing and perceiving errors made up 24% of all accidents, and incapacitation made for 10%. These particular crashes are thought to have been potentially avoided if all on-road vehicles were self-driving. However, these autonomous vehicles would need to have zero malfunctions, and the remaining crashes could still take place unless these vehicles were able to specifically avoid other decision-making errors.
Duke University robotics and human factors professor, Missy Cummings, gave her thoughts on the study. She explained that technology is not likely to prevent even one-third of crashes caused by human error right now, as self-driving vehicles with radar, laser, and camera sensors still often cannot perform perfectly in any situation.
“There is a probability that even when all three sensor systems come to bear, that obstacles can be missed,” she said. “No driverless car company has been able to do that reliably. They know that, too.”
Cummings also explained that those working in the business of self-driving vehicles did not plan for any technology to prevent all human-caused crashes. That belief, she said, would be “layman’s conventional wisdom that somehow this technology is going to be a panacea that is going to prevent all death.”
Researchers from IIHS who studied crash causes made the ultimate decisions about which ones could be prevented altogether with only autonomous vehicles on the road, according to Cicchino. When self-driving vehicles share the road with human drivers, even fewer crashes will be able to be prevented.
“Our analysis shows that it will be crucial for designers to prioritize safety over rider preferences if autonomous vehicles are to live up to their promise to be safer than human drivers,” said Mueller.