How Bosch is preparing for autonomous cars

Things were very different when Gerhard Steiger started his career at Robert Bosch three decades ago. Failure was taboo. However, there has been a big shift in mindset at the world's largest supplier since Steiger took over as the head of chassis systems five years ago. When it comes to finding groundbreaking solutions that will benefit the forthcoming wave of fully autonomous vehicles, failure is encouraged as long as it happens quickly and it provides a valuable lesson. Steiger discussed how Bosch is adapting during an interview with Automotive News Europe Managing Editor Douglas A. Bolduc.

How is the move to autonomous driving changing things at Bosch?

At Bosch, we have a dual approach. On one hand, we foresee a steady increase in partially automated driving, going from Level 1, to Level 2 to Level 3. This follows the classical business model. On the other hand, we are taking a non-traditional, more Silicon Valley-like approach for our work on the shared autonomous vehicle, or robotaxi [which will be a Level 5 driverless vehicle].

This dual approach is new for Bosch. Why did it need to change?

In the past, failure was never an option, especially for us as a safety systems provider. That is still the case, but in some instances -- long before the technology or product ever enters a vehicle -- failure is an option, as long as we can learn from it to create something even better. This change in mindset is the result of Bosch's decision to encourage more of a startup mentality within the company. This was done to let new ideas develop.

How do you find the right balance?

As head of the chassis systems control division, my job is to make sure that there is never a failure in our traditional business because no one would accept it if their ABS or stability control failed. But as we look for new opportunities I sometimes need to take risks on new ideas.

Bosch and Daimler are co-developing fully automated, driverless vehicles for launch at the start of the next decade. What role will each play?

We will have joint engineering so we can co-develop everything related to perception, trajectory planning, safety and more. Also, all software will be done jointly.

Will Bosch be doing anything by itself?

Bosch is developing the sensors.

Will you sell end products to others?

Although we decided to co-develop with Daimler, we will also offer them to others. It could be traditional automakers. It could be all-new customers.

Why did Bosch need to adapt even in an area where it excels, automotive?

Because the environment is so much more complex it requires a different level of technology, including more powerful ECUs [electronic control units] and sensors. There is also a greater need for redundancy when you have three different sensing technologies -- the camera, radar and lidar -- monitoring 360 degrees around the car. If the camera fails, you still must have enough sensors to get a good view of the whole environment.

How is the business model of your non-traditional arm different?

You are no longer selling cars to individuals. You're selling them to fleets. The benefit comes from the much higher usage of the car because there is no human involved. Since it's driverless, you save this cost, which is a big portion of ride-sharing today. That allows you to put more technology in the car. This will definitely make the car more expensive than what you see now.

What are the next products that will come from the more traditional side of your business?

We will launch our traffic jam pilot [a system that accelerates, brakes and steers the car at speeds up to 60 kph] between 2018 and 2019. The next step after that is our highway pilot, which will take over on a motorway with the push of a button. This is going to come at the beginning of the next decade.

What is the biggest hurdle in the move to fully autonomous driving?

The biggest challenge is making sure that this will be really reliable and safe. It is crucial that everything related to autonomous driving is safe. Many players underestimate the effort that it takes to make it bulletproof.