90 per cent automation, 10 per cent human

What is Fernride’s objective?

Together with strong partners, we want to offer solutions that create the transformation from ‘diesel and manual’ to ‘electric and autonomous driving’. Even though Fernride has only existed for four years, we are not that far away from this.

Autonomous driving has been discussed for a long time – but at present, how realistic is it to expect it to be brought onto the road in the near future?

There are still regulatory obstacles to overcome, but realistically, over the next few years we will take the next step and use mixed traffic in complex environments – mobile robots that virtually think like a human. But this also means that we are not going to start where the hype about autonomous driving might have started five to ten years ago, when people were dreaming of “robo-taxis” and designing “hub-to-hub” transports that involve very complex technical problems. I am sure these problems will also be solved in the next 20 or 30 years, but we at Fernride have asked ourselves how we can make a start today and generate added value.

How are you going about it?

When we founded Fernride in 2019, we had ten years of research experience behind us in the field of automated driving at the Technical University of Munich. There, we primarily investigated teleoperated driving and this has strongly inspired our product vision and technology. We are convinced that it is not necessary to drive one hundred per cent autonomously, rather we use as much autonomy as can now already function. There are certain gaps, so-called edge cases, that concern e.g. safety – these add a lot of complexity. And it is precisely these that we do not leave to technology, but make the responsibility of people. But the driver controls the vehicles remotely: from the control centre, which can be a simple office where they are operating a console. They can even operate several trucks from there at the same time. Today this is just four, for the future we are planning up to 50 and more. Then the trucks behave like mobile robots that can be used to automate the work processes of logistics companies.

How much autonomy will be possible in the future?

A huge amount. At the moment we are using human power even more, for example for direct control at a construction site that has just been rebuilt. The algorithm cannot replace them right away. But when it has seen this 1,000 times, it can learn from the data. And hence the level of autonomy can be steadily increased, more and more in the direction of 90 per cent.

And the missing 10 per cent?

This should not be automated at all, as far as we are concerned. Instead, we want man and machine to work in harmony. Because that removes complexity and hence also development time. Every work process requires other human skills, we call them secondary activities: driving from A to B is the baseline. But drivers are also responsible for coupling and uncoupling, opening and closing doors and securing loads. And some of their activities should not be automated because that would be much more costly than if people continued to do it.

Will truck drivers nevertheless soon be superfluous?

No, in fact we can create attractive jobs in this way. The driver shortage is acute, there is a lack of prospective new drivers. But at the same time, all those who are doing this job today must be kept in the loop. After all, it is not just about technology, but also the question of how to make a cultural change successful. For example. you can’t just turn people’s lives upside down and tell them that their jobs will no longer be there tomorrow. Instead, it is important that they know: “your jobs will be more secure!” And you will climb out of the cab and into an office space where you can do higher-level work and control many vehicles.

You are already showing in practice what this looks like. Whereabouts are you active?

At Volkswagen in Wolfsburg and at DB Schenker distribution centres, among other places. We want to gain experience in live operation there, in order to take the next steps based on that: away from the test track and onto the street. But there are also many challenges on the demarcated site: there are other trucks driving in and out, or pedestrians walking around. There are many situations that are familiar from inner cities, and these give us valuable impulses. We are setting up a roadmap with our partners to get us to autonomous electric truck driving within the next ten years.

One of these partners is HHLA Next, a start-up that functions as the innovation and “venture building and investment” unit of Hamburger Hafen und Logistik Aktiengesellschaft (HHLA). What exactly are you implementing together?

In Tallinn, Estonia, HHLA International operates a terminal where driverless trucks are already moving containers in live operation every day. The port handles heavy duty stuff – large masses of 20 to 45 tonnes. This is a correspondingly high-risk workplace, but the processes are predestined for autonomous driving. The project in Tallinn is no longer a laboratory, but a place where real service level agreements are operated with our vehicles in real time.

How will it continue?

The next step for us as an industry is to prove that this is no longer a research project, but that autonomous vehicles can function in live operation – not just with one customer, but with several. And then also to scale it jointly for the first use cases. After all, we are in logistics, and this must bring added value. It’s not just cool to ride in a robo-taxi, we also want to save money and realise economic benefits as well as qualitative advantages such as sustainability and safety. In doing so, we want to change the existing infrastructure and interfere with logistical processes as little as possible.

What still needs to happen for autonomous driving to become established?

What certainly does not stand in our way is political will or a desire on the part of the customers. Instead, we have to design the technology very precisely and develop a technology that is safe enough. To do this, we have to precisely analyse the scenarios that occur in reality and break them down into individual technical systems. Every tiny component must be designed to be as fail-safe as possible, and the probability of an accident is further minimised by redundantly integrating these components. This requires not only software, but also hardware: powerful brakes are needed to bring a fully loaded truck travelling at 60 km/h to a halt within eight metres in snowy conditions. So we also need changes and product maturation from our suppliers and system integrators that are not yet here today, but will arrive in the next few years and then also go into series production. That will be the great enabler to get this on the road.