IIHS Study: Autonomous Cars Won't Avoid Majority of Vehicle Crashes

METAMORWORKS - GETTY IMAGES

Meanwhile, autonomous-vehicle developers find issue with the study and its conclusions.

• An IIHS study determines that autonomous cars would still struggle to avoid two-thirds of crashes.
• While self-driving cars eliminate issues around detection and incapacitation, if they are built to be more aggressive than safe, they’ll still end up crashing.
• The IIHS concludes that self-driving vehicles "would need to be specifically programmed to prioritize safety over speed and convenience" Something AV developers say they are already doing.

The general consensus has been that once autonomous cars take over the roads, car accidents will almost be a thing of the past. But according to a new study from the Insurance Institute for Highway Safety (IIHS), not only is that not realistic, but it’s likely that self-driving cars might only prevent a third of crashes if the systems are set up to drive the way people drive now.

The IIHS concluded that, while self-driving vehicles would be better at detecting the world around them thanks to a suite of sensors that produce a 360-degree worldview, they'll still get into crashes. That's even if all the vehicles on the road were autonomous.

The IIHS reviewed more than 5000 police-reported crashes from the National Motor Vehicle Crash Causation Survey. The collisions must have resulted in at least one vehicle being towed away and emergency medical services called to the scene. They split the factors that contributed to the crashes into five categories:

• "Sensing and perceiving" errors included things like driver distraction, impeded visibility and failing to recognize hazards before it was too late.
• "Predicting" errors occurred when drivers misjudged a gap in traffic, incorrectly estimated how fast another vehicle was going, or made an incorrect assumption about what another road user was going to do.
• "Planning and deciding" errors included driving too fast or too slow for road conditions, driving aggressively, or leaving too little following distance from the vehicle ahead.
• "Execution and performance" errors included inadequate or incorrect evasive maneuvers, overcompensation, and other mistakes in controlling the vehicle.
• "Incapacitation" involved impairment due to alcohol or drug use, medical problems, or falling asleep at the wheel.

The insurance nonprofit deduced that 2 percent of crashes are unavoidable and 4 percent are from unknown causes. Apart from those, the group determined that sensing and perception error collisions account for about 24 percent of crashes and incapacitation for an additional 10 percent. Those would essentially be wiped out by vehicles that can see the world better than us and won't get drunk. In other words, autonomous vehicles could cut crashes by one-third, potentially.

The remaining 60 percent, according to the IIHS, really rely on those building the systems and the person sitting in the car. The study found that 40 percent of crashes were caused by planning and deciding errors like speeding and illegal maneuvers. The study states, "The fact that deliberate decisions made by drivers can lead to crashes indicates that rider preferences might sometimes conflict with the safety priorities of autonomous vehicles."

In other words, if costly autonomous vehicles are built to be as aggressive on the road as human drivers, crashes will continue. "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," IIHS research scientist Alexandra Mueller said in a statement about the study.

But not everyone agrees with the studies findings or how the IIHS came to its conclusion.

The Alliance for Automotive Innovation, a organization dedicated to helping innovation come to market safely and cleanly that counts BMW, Toyota, Ford, GM, Honda, and others as members, sent Car and Driver the following statement: "This study confirms the life-saving value of advanced vehicle technologies, but we believe that reducing traffic fatalities by even a third would be something to be proud of. We aim to do even more."

And it's true that reducing crashes by as much as a third would be huge. According to an NHTSA early estimate, 36,120 died from motor vehicle crashes in 2019. If that could be reduced, that’s an impressive feat.

The Alliance continued, "We all know this about autonomous vehicles: they hold tremendous promise in saving thousands of lives on our roads every year, and the promise will be even greater as the technology matures. Aside from saving thousands of lives annually, AVs can help usher in a new era of mobility for those currently limited due to physical challenges, as well as contactless movement and delivery, which is more important now than ever."

The Partners for Automated Vehicle Education (PAVE) a coalition of partners created to help educated policy makers and the public about autonomous vehicles also finds issue with the study: "The IIHS study helps provide better insight into the cause of road crashes, but falls short in its application of these insights to autonomous vehicles by assuming that superior perception and the ability to avoid incapacitation are the only ways in which autonomous vehicles can improve on human driving performance. This is simply not the case."

PAVE continues, "one of the great advantages of autonomous vehicles is that their behavior can be programmed to never speed or break the laws, behaviors that this study blames for 38-percent of crashes. The assumption that these behaviors can be altered by passengers in ways that so drastically reduce safety is inconsistent with what our members tell us about the culture they bring to AV development, as well as the goals and practices of their development programs."

Jack Weast, vice president of autonomous vehicle standards at Intel's Mobileye, told Automotive News, "crashes will never be zero until we have no more human drivers on the road," he said. "But (self-driving cars) can combine physical laws with behavioral studies and do much more than a human driver."

Still the IIHS has its reservations. "Building self-driving cars that drive as well as people do is a big challenge in itself,” Mueller said. “But they’d actually need to be better than that to deliver on the promises we’ve all heard.”

SOURCE: CAR AND DRIVER