We have all heard about the new self-driving vehicles, which Google has put into the spotlight within the past few years, and other companies are now getting onto the self-driving vehicle bandwagon. There was a lot of selling points to these self-driving vehicles, including that the thought was they were much safer than your traditional car with a person driving. This is because real people get distracted texting or talking on the phone, and self-driving cars eliminate someone getting behind the wheel drunk.
The self-driving cars have a bunch of sensors on them, which gives them a 360 degree view of their surroundings, so it would make sense that these are the cars of the near future. While the previous thought was that self-driving cars are safer, a new study that was published from the University of Michigan is showing the exact opposite in reality, meaning that self-driving cars are getting into more accidents.
The new University of Michigan study is saying that self-driving cars are getting into more accidents per million miles traveled than the traditional cars out there. Interestingly enough, all of the accidents up to this point with the self-driving cars, have been the fault of the human driving the other car. The University of Michigan was able to do this study because self-driving cars have been on the road long enough to get a good representation of the data about these cars, and we can now look into the real world performance of these self-driving cars. It seems almost perfect to think about the self-driving cars getting into more accidents on the road than the traditional and conventional cars, given that these accidents are the fault of human error in the other car. There are quite a few important notes to go over too in the new findings, and we will tell you about some of these caveats to the research data.
While the self-driving cars have been on the roadways for long enough to get some good data, the self-driving cars have only logged over 1,000,000 miles, and this is still not enough miles to give a large enough sample of autonomous vehicle accidents. Throughout the 1 million miles autonomously, there have been 11 crashes while in the self-driving mode. This is a very small sample so it is hard to get hard statistics or draw concrete conclusions from it.
There is also a little uncertainty when it comes to how many accidents the conventional cars get into, since sometimes drivers do not report the accidents if there is minor damage to one or both of the cars. The Department of Transportation does estimate that there are a specific number of unreported accidents to try to adjust for that but those numbers might be really off compared to the real numbers of people who do not report accidents. When you look at the self-driving cars, the companies have to report each and every accident, thus it makes it easier to get a concrete number and very accurate data.
Another issue is that self-driving cars do not end up in all of the challenging situations that real human drivers have to deal with while on the roadways. In this particular study, the data all came from California and Texas and those similar states, which have milder winters compared to the states like New York and Pennsylvania where you have a ton of snow and ice to deal with in the winter. There is also the issue of the self-driving cars likely not being used in the same high-traffic situations that real drivers have to deal with, which makes it less likely they will get into accidents.
Lastly, the self-driving car accidents are also less severe even though the self-driving cars are getting into more accidents on average. When you look at the self-driving cars, only 2 out of 11 had injuries related to the accidents, whereas 28 percent of conventional car accidents led to injuries. A lot of the self-driving accidents occurred as a result of another car rear-ending the self-driving car, and no head-on collisions have been reported with the self-driving car at all.