ZF exec looks for help to reduce traffic deaths
AUTOMOTIVE NEWS EUROPE MONTHLY MAGAZINE
Nearly 60 companies looking to help ZF create next-generation solutions for autonomous driving, connectivity and artificial intelligence had 3 minutes to present their ideas at the supplier’s so-called “pitch night.”
There is a big number that ZF Friedrichshafen’s head of advanced engineering is determined to reduce. That number is the 1.25 million traffic deaths that take place worldwide each year. Torsten Gollewski, a former Audi manager who joined ZF last year, is not alone.
Rival engineers at suppliers such as Continental, Delphi and Robert Bosch are racing to find solutions to help prevent this massive loss of life. To get there, Gollewski, who also serves as managing director of ZF’s Zukunft (German for future) Ventures incubator, is looking for help.
Zukunft Ventures acquires stakes in small companies that help bolster ZF’s expertise in autonomous driving. One example is the 45 percent share that the world’s second-largest supplier took in Astyx Communication & Sensors earlier this year. ZF has teamed with Astyx, a maker of ultrahigh-frequency radar sensors, to develop algorithms that can predict things such as the frictional coefficient for a curve on a deserted country road.
In Germany, roughly half of all traffic-related deaths in passenger cars occur in single driver situations, such as understeering on a slippery road. Here, for example, even the most modern car-to-X communication system would not be able to provide a vehicle’s self-driving computer with up-to-date information about surface conditions, since it might have been an hour or more since the last car drove by and transmitted a measurement to the Cloud.
ZF’s Gollewski wants next-generation models to know what to expect around the next corner.
“We have to know what to expect around the next corner – is the asphalt slick, is it covered by snow, is it dry,” Gollewski told Automotive News Europe.
To strengthen its expertise in sensor data, ZF also acquired a 40 percent stake in Ibeo, which is developing a solid-state lidar capable of mapping its environment using laser beams emitted over four vertical levels, or layers, to give a three-dimensional picture. Larger rival Robert Bosch, by comparison, is still evaluating whether to enter the market.
Each stake-taking decision is discussed by the Zukunft Ventures board, which also agrees on the financial resources to be invested, giving the fund a flexible capital base. “If you look at other companies, most have fixed amounts of capital, anywhere between 100 million to 500 million euros depending, and invest smaller amounts in a number of companies to diversify their risk,” he said. “That’s a big difference to our approach, since we want Zukunft Ventures to help ZF implement its strategy.”
When it comes to acquiring strategic know-how, Gollewski says ZF is willing to move briskly “We have not been limited internally in any way thus far when it comes to striking a deal,” he said. “If you need a decision, you’ll get one.”
While ZF has been acquiring expertise in the field of sensory perception via Zukunft Ventures, the second part of the equation – decision-making by the vehicle – is a puzzle piece that the hasn’t yet been found. An analysis of crash data, for example, shows that most accidents could have been prevented had the car changed trajectory 500 milliseconds before the collision. The challenge is that an autonomous vehicle needs a lot of computing power to fuse sensor data and then react within a fraction of a second.
While the ZF executive says he would rather measure his performance by the quality of the investments rather than the quantity, Gollewski said his team continues to screen candidates from around the world, including companies in the area of artificial intelligence (AI). To promote a closer cooperation with promising entrepreneurs in the area of autonomous driving and AI, the supplier hosted a so-called “pitch night” with 57 startups from 17 different countries at its headquarters in Friedrichshafen, Germany, earlier this year.
However, until the time when nature invents a “Deer 4.0,” as Gollewski puts it, there will always be variables during autonomous driving for which an algorithm will be unable to deal with and the vision of zero traffic deaths will remain an honorable, if unreachable, goal.
You can reach Christiaan Hetzner at email@example.com.
Let’s block ads! (Why?)