Michael Fausten

Michael Fausten

How will tomorrow's car drive and what role will you play?  Those in favour of autonomous driving are quick to point out the benefits, such as fewer accidents, reduced traffic congestion, improved fuel efficiency and, of course, relieving you the task of driving, parking and navigating the vehicle (something we all might especially appreciate as we live for longer). But not everyone likes the idea of driverless cars running around on our roads. A sudden malfunction could leave you powerless. For its part, Bosch used this year's IAA to display an avalanche of innovations to support autonomous driving.  In this interview, Matthew Beecham talked with Michael Fausten, Head of Bosch project team Autonomated Driving.

Some OEMs are expecting the autonomous car to appear as early as 2020. What's your view?

We both share this vision and follow it. We see two steps on how to introduce automated driving technology. The first step is through what we call partial automation, i.e. the vehicle does the driving but the driver is obliged to supervise what the vehicle is actually doing. The second step is the high automation up to full automation where the 'driver' could read the newspaper, text or anything but actually drive the vehicle.

Partial automation is coming soon.  For example, a Bosch traffic jam assistant will go into series production next year, and we will push that towards free driving up to perhaps 130kph by the middle of the decade. As for full automation, we share the vision of that happening by 2020 even to that speed.

Technology is one thing, regulations are another but consumer acceptance of such technologies needs to happen too.

We see customer acceptance as surprisingly good. People like to have the freedom of choice to say: "I want to drive on my own," or, "I want my car to be doing the job for me," and that's the way we have to project it. If you want your car taking over the task of driving thereby giving you time to dedicate to things you would like to do, then you would probably be highly interested in getting this. Placing the question that way, we have found that more than 50% of people agree to this technology.

There are some time honoured arguments that we hear liability issues if the technology malfunctions. What's your view?

For partial automation, the situation is clear; you are obliged to review what the vehicle is doing and take action if you feel that something is going in the wrong direction. That is the same way as it works in driver assistance too; if you are driving an ACC system you can't just crash into the rear end of the car in front of you and say, "Hey, the ACC didn't take action." It's your obligation to step in, the same way as with the partial automated systems.

For the highly automated system, the situation changes completely. The driver will be permitted to go out of the loop and out of control. Let's take legislative regulations for granted in that case, because some aspects still have to be modified there. Then yes, the driver is out of responsibility. Certainly, we would not put a product into the market if we are not sure that it is safe. So we must test the systems, review and regulate them until we really think they will behave according to the rules of traffic and they will be capable of handling all the relevant cases. But yes, if an accident happens, you need to have other methods how to check whose guilt is it. One solution might be recording the driving data and what the car sensors captured all around. By drawing these figures, you can say, "It was not the automatic car, it was just the other guy committed the fault."

Are we giving the driver too much assistance today?

At least we have to think about how we give this assistance. We are paying careful attention to how systems dates could change from one to each other, so how do you engage a partial automated system? How should the vehicle behave in case it disengages? How should we make the handover back to the driver? How shall we present the information to the driver that the vehicle is in need of intervention? How much time will it take for the driver to take over to capture the scenario and then really take action? Up to the point to say, "How do we check that the driver is, in the case of partial automation, still aware of what is happening?" So it's probably not the question of whether we are giving too much assistance but rather about how do we interact with the driver? How do we interact with the human being? What are the capacities of the human being? What are his limitations? How do we interact and how do we keep him in the loop in case we need him? So it's rather that question that they are going after than the question of whether it's too much.

In terms of its autonomous vehicle technologies, what is the message that Bosch would like to send from this year's IAA?

The key message is that Bosch is taking this technology seriously. We believe that we will enter the market and we are deeply involving our engineering forces into understanding the system and the system's requirements so that Bosch will be in the position to provide system know how as well as the right components at the right time.  Especially highly automated driving will involve lots of changes to the vehicle architecture, it will put very high requirements to control units, bus systems, sensors, braking systems, and you really need to understand the whole system in order to have the right products in place.  That is what Bosch is doing.

Is there an appetite for this ADAS technology in China?

There is interest for driver assistance systems and as well for advanced driver assistance systems. If you take a glance at the news you will find that one of the other, there's the OEM who is really investigating together with universities or institutes how automation would work. So we expect that as well as China we have some interest here.

To what extent has traffic sign recognition software advanced?

The remainder of this interview is available on just-auto's QUBE research service

Auto market intelligence
from just-auto

• Auto component fitment forecasts
• OEM & tier 1 profiles & factory finder
• Analysis of 30+ auto technologies & more