by Nicolas Simon,
Leonard Storch, Germany
Think of the self-driving car, an idea that seemed impossible 10 years ago and yet it’s already a reality on some highways. It is not like they are regulary on the streets, but although some of them are being used by companies like Tesla or Uber. These companies are seeing like many others the future of traffic and cars in these autonomous vehicles. So, in case you are asking yourself why these cars aren’t used much more on a daily basis, it is because with every new invention comes also a new danger. Even though the vehicles have an impressive safety record compared to the carnage wrought by human-driven vehicles, their excellent performance is largely enabled by machine-learning algorithms that have been trained on torrents of data about roads, intersections, street furniture, etc. But sometimes there have been fatalities. In 2016, for example, a self-driving Tesla drove into the back of a white tractor-trailer; in March this year, a self-driving Uber car killed a woman pushing a bicycle and so on. These accidents are called “edge cases” because the cars’ software encountered scenes or objects that it didn’t recognize for the simple reason that they had never appeared in the software’s training datasets. Given that the real world is full of things that cars’ sensors and software have never seen before, these edge cases will continue.
So in the end I would say that there is certainly a need, and many benefits, for autonomous cars, but on the other hand, the current technology doesn’t quite seem ready for full autonomy all the time.