Pedal to the metal: changing course toward autonomous driving at full speed
Electric, connected and automated: technological trends are driving carmakers to fundamentally rethink their organizational structures. Their conventional approach is unlikely to lead to fully autonomous driving anytime soon. That is why German auto companies are now switching to the fast lane when it comes to R&D.
There’s a lot of back-patting going on in Germany, the country that invented the automobile – for obvious reasons: nearly 72 percent of all premium vehicles used worldwide are German brands. Almost half of the turnover of the German automotive companies stems from innovation – an unparalleled figure in the industry. When it comes to the development of driverless cars, German automakers also seem to have the edge: according to a new study by the Institute of the German Economy in Cologne (IW), Germany accounts for 58 percent of the worldwide patents filed under this field since 2010.
On the surface, it looks like German car makers are on top of the game.
However, all is not what it seems. The problem with the IW patent rankings is that it only counts patents for which the applicant has mentioned autonomous vehicles directly. Patents pertaining to general technologies in the fields of artificial intelligence and image processing for example are not considered. For that reason, Google sits in a mere 5th place in the rankings for autonomous driving patents per se – while suppliers Bosch (1st place: 545 registrations) and Continental (3rd place: 277 patents) occupy two spots of an all-German podium. They are joined by the car manufacturer Audi (2nd place: 292) while General Motors just missed out (4th place: 246). Honorable mentions go to Volkswagen (184), Toyota (164), Daimler (156), BMW (142) and Ford (103).
The numbers, however, do not fully reflect reality: with the growing importance of A.I. and self-learning technologies comes a shift in traditional power structures. "New competitors are entering the market and changing the rules radically," warns Hubertus Bardt, author of the IW study. Although the study found that only around 7 percent of worldwide patents for autonomous driving are submitted by Google, Apple, Tesla and Co., it is they who "accelerate the competition," according to Bardt. "With business models that focus on digital components, networking technologies and Big Data, these companies find themselves in an extremely strong position,” he adds. One could even go as far as to say that they are the technology leaders and it is the traditional car manufacturers and suppliers that have the catching up to do.
Reaching Level 5: evolution vs revolution
Until not so long ago, many of the auto industry’s familiar players chose the evolutionary course of development when it came to automated driving. This involves adding to and improving the many separate sensor modules that give rise to driver assistance systems. However, no amount of evolution in this sense will provide the car with a complete and comprehensive image and understanding of its surroundings. That requires a revolution.
Only if the system is given a “brain” – a higher decision-making body – will it be able to truly master the flood of data: filter it, interpret it, develop a driving strategy and even make ethical decisions if an emergency situation calls for it. Therein lies the difference between reaching Level 4, where the automated car can master only defined use cases (e.g. highway driving) without the driver as backup, and Level 5, where the system is able to act alone in all use cases. If faced with complex inner city scenarios involving pedestrians or cyclists, systems developed in an evolutionary manner are hopelessly overwhelmed.
In the meantime, Google has been working furiously on a central computer model for highly automated driving. The IT giant’s fleet of autonomous test cars has already clocked several million kilometers since 2009 – far more than any automaker. With all that data collected they can develop algorithms and neural networks that are helping robocars make the right decisions.
BMW’s goal: full automation by 2021
So as not to fall behind Google, many OEMs have started to ring the changes – and BMW is no different. The Bavarian auto-maker’s boss, Harald Krüger, has triggered a radical strategic reform that will see the company through the next decade. ‘Fully automated driving and driver assistance systems’ is the name of the new mega-department for vehicle development. The new unit comprises, as of the end of last year, E/E specialists, chassis engineers and A.I. experts all working together toward one goal: led by Elmar Frickenstein the integrated team aims to bring the fully automated BMWi Next into series production as early as 2021. Their recent recruitment reflects their dedication – having hired 400 IT professionals last year with another 500 experts in artificial intelligence to be added this year.
"Automated driving will revolutionize the automotive industry" says Elmar Frickenstein. Indeed, the BMW i 2.0 project focuses on both pre and series development: the goal is to massively gain speed in getting ideas to market. "We’ve adopted a two-pronged approach. In parallel to the evolution of existing vehicle architectures and driver assistance systems, we are taking a revolutionary approach in developing future generations of connectivity and cloud-based services," says BMW Chief Development Officer Klaus Fröhlich. The goal is to be the leader in autonomous driving technology: "This is an enormous challenge. But we will make autonomous driving available when others have not yet dared to do so due to financial or other reasons," vows Fröhlich.
Not to be outdone: Mercedes invest in the revolution
Mercedes has also doubled down in this sense. Under the direction of Dr. Michael Hafner, existing driver assistance systems and active safety will continue to be developed. However, he acknowledges the limitations: "Road traffic has become so complex that many of our systems cannot be made 100% available. The final percentages are the most difficult - and we don’t want to introduce a system that could leave the driver alone in an emergency from one second to the next," said Hafner. At the beginning of this year, Mercedes received a license for autonomous test-drives with its new E-Class in the state of Nevada. For testing only software modifications in the DRIVE PILOT electronic control unit are needed. Additional sensors, modified steering, an adapted ESP or a super computer in the trunk are not required.
But is this basic equipment enough for driverless urban robo-taxis? That question is left to another team at Mercedes, led by Prof. Dr. Ralf Guido Herrtwich. In a similar two-pronged approach, some 200 experts are working in parallel to the production vehicle evolution – with a view to revolution. In collaboration with the University of Ulm, the Karlsruhe Institute of Technology (KIT) and FZI Forschungszentrum Informatik, Mercedes have founded the Tech Center ‘a-drive’.
Here, the integrated team is working on groundbreaking technologies like trajectory planning: ensuring that the autonomous car is not only collision-free and legally compliant, but also safe, efficient and cooperative in the sense of feeling natural to the human occupant. However, it was not only Daimler who invested in ‘a-drive’. The tech center is a 7.5 million Euro project of which Daimler AG contributed 5 million with 1.25 million coming from both the Ministry of Science, Research and Arts and the Ministry of Finance and Economy in the state of Baden-Württemberg.
Whether or not the German OEMs like BMW and Mercedes can prevail against Google and Uber remains to be seen. What can most certainly be seen however, is that the traditional automakers are taking action to stay in the race. The automated wheels are in motion!
- While German premium manufacturers fare well when it comes to patents for automated driving, newcomers and Silicon Valley companies go straight for the revolution towards autonomous vehicles.
- This disruptive branch of development triggers organizational changes in the automotive industry.
- BMW created the mega-department “Fully automated driving and drivers’ assistance systems” and the project “BMW i 2.0”.
- Mercedes Benz founded the Tech Center “a-drive” to work directly towards the urban robo-taxi.
Who do you think will win the race to autonomous driving? Is it too little too late to catch up with the revolutionary companies such as Google?