Project Manager Brandon Schoettle said that self-driving cars have some major barriers to overcome before they can be considered safer than vehicles driven by experienced, middle-aged drivers.
“That zero fatality level, based on certain things we’ve seen, isn’t likely to occur because there are other things than driver error that go into fatal accidents,” Schoettle said. “It’s possible that in smaller countries, maybe the Netherlands or Sweden, it may be possible for them to get very close to, if not maybe hit, zero [fatalities] occasionally. In a country as big as the United States, there are just random events that happen [that keep that figure above zero].”
The report points out that many scenarios that cause fatalities are not the result of driver error. Jaywalking pedestrians, mechanical failure, and damaged roads and infrastructure can all cause problems that in some cases are more problematic for self-driving vehicles than they are for human drivers.
While self-driving vehicles level the playing field for human passengers of all ages, Schoettle said, the level of safety the technology can achieve is not yet known. They hope that at the very minimum the technology will enable vehicles to drive more safely than even the safest human drivers and ideally much more safely than that, but that conclusion is not foregone, he said.
Things might also be complicated further during the overlap period between a future of all self-driving vehicles and today where human drivers must share the road with self-driving vehicles, and expected-but-missing driver-to-driver interaction could create confusion on the road.
“We’re at this strange point with self-driving vehicles,” Schoettle said, “where there are certain things they currently do much better than any human driver ever has – their ability to constantly watch the road, to adjust to really small changes, to maximize use of the braking system and other performance factors of the car is much better than a human driver even at this point, but they’re still easily confused and have to shut down in certain situations where a human driver is able to quickly think through it as a part of everyday life.”
In May 2014, UMTRI began construction of the Michigan Mobility Transformation Center, a $30 million simulated city that will help researchers test self-driving and interconnected vehicles in an environment simulating the real world. The 32-acre facility was primarily funded by the U.S. Department of Transportation, and research partners like Bosch, Econolite, Ford, General Motors, Toyota and Xerox expect to begin testing their products in the facility upon its scheduled opening this spring.
