Interview with Aurora co-founder Chris Urmson

'Aurora' is the Latin name for 'dawn' - that is, the reddish color of the sky before Sunrise. We found the name to be quite appropriate for the company, because the name literally stands for a kind of “new beginning”. Autonomous driving stands for the transformation and revolution of the automotive industry and mobility. The meaning will be fundamental. And we're still at the very beginning.

True, there is a lot of talk about expensive laser technology and written. But if you look into the laser and look at the components, there is nothing exotic in it. The high price is actually due to the fact that very few have been produced to date. The costs in relation to the volume simply remain very high.

We try to do so to be as quick as possible. We believe that we have not only recognized and analyzed the great challenges of autonomous driving, but also understood them. But we are very reluctant to attach a date to our project.

Aurora sends its systems in one Lincoln MKZ for testing on public roads.

Aurora is focusing on developing a kind of “artificial driver” for future vehicles. Such systems are not yet on the market - this technology is only just beginning to move towards series production. Aurora makes it its business to make autonomous driving as fast as possible and as safe as possible for youdevelop wide market. We say: let's do what we may be best at in the world - build a safe vehicle chauffeur. We can thus give many people the mobility they appreciate. Nobody likes to be stuck in traffic in the morning on their way to work. So let's make this part more enjoyable and safer than ever before. Aurora wants to develop an 'artificial driver'. That will take a few years.

Yes, absolutely correct. There is a market for the system and the artificial driver, and it is quite attractive.

Let's take a VW Golf. You have a prototype that is likely to cost 100,000 euros each. That's still like that. But these prototypes are unique and far too expensive to produce thousands of times. And it's similar with us: We're also at the prototype stage; the technology cannot go into series production like this yet, but it is on the verge of it. We expect that the costs in mass production will go down considerably when we enter the volume business.

We want to develop this 'artificial driver' that our customers can use wherever they find it useful and valuable. Some might prefer to use it in shuttle services or taxi services. For other customers, it may be more interesting to install it in vehicles for private use so that the owner then has a chauffeur to drive them around. We're sticking to what we do best: developing this artificial driver. We then look together with our customers - for example Volkswagen - how we can best use the system and develop visions.

Absolutely. We agree.

We cannot say how ours are Technology is ultimately used. The next few years will show in which vehicles, means of transport or in which contexts this technology is used. And also in which cases it pays off. Our customers will also have to find out first.

In my opinion, the value of testing is not in the number of kilometers, but in identifying the things and situations that lead to an error. We do not learn from the number of successful kilometers, but from the many mistakes and their causes during the test drives. The most important thing for us is how we can quickly identify the problem behind an error in order to be able to fix it. We are therefore rather looking for environments where the system does not find its way so easily.

We are still at the beginning of this technology and are already making enormous development steps. We test all possible situations first in the simulation and only then on the road. With the test drives, the greatest challenge is predicting the behavior of other road users. That really is the greatest challenge of all. Most of the time, it's not that difficult to define how other people behave in traffic: they mostly stay in their lane, passers-by mostly stay on the sidewalk, and if everyone obeys the rules, things go pretty well. But if that is not the case, then it will be difficult to make a prediction. Because every action can also lead to a reaction from third parties. This is the greatest difficulty in giving the system follow-up rules that really apply to the situation.

2017 shows VW with the Sedric study, an autonomously driving vehicle the size of a VW bus.

Our goal is to make the technology ready for the market. Completely independent of the respective partner. So we are working with our partners on concepts and plans on how to use and integrate the technology in their interests.

We work together with Volkswagen. VW has a fairly large portfolio and therefore theoretically many possible uses - even up toTrucks. After the Dieselgate, they put a lot of thought into where to go with the company. It was a bad time for the company and bad things happened. But that made them define where they want to go with a much higher intensity. We had the opportunity to talk to them at this time. That was really exciting.

We hope Aurora is one of them five companies. But it is really too early to name the remaining competitors. When I meet and talk to my former colleagues from Google, they have really gained a lot of deep experience on the subject. And there are also some automakers who have come a long way.