The automotive industry is at the forefront of technological progress, with artificial intelligence (AI) as a cornerstone technology. Continental, a leading global automotive supplier but also a tech company, has recognized this trend and opened its own AI lab at the end of 2021. Here, a team of experts is working on groundbreaking AI technologies with the aim of shaping the future of mobility.
At the heart of this innovation is Dr. Andreas Weinlich, Head of the AI Lab at Continental in Berlin. In our interview, he shares his experiences, gives us a direct insight into the day-to-day work of his team and shows how they are exploring the possibilities of AI to change the mobility landscape.
Dr. Weinlich, Continental is known worldwide for its leading role in the automotive industry. With the opening of an AI lab in Berlin, you are expanding your capacities in the field of artificial intelligence. Could you tell us more about the motivation behind this decision and why Berlin in particular was chosen as the location?
AI is becoming increasingly important within our company as well as in the automotive industry as a whole. In this incredibly fast-developing field, it is important to be able to jump on the latest technologies quickly and this is only possible if you are close to both research and the very agile start-up scene. We have achieved this through our offices in the AI Campus, as a co-working space with daily events on the topic as well as excellent networking opportunities. By cooperating with selected universities, we also guarantee efficient technology transfer to our company.
You work with technologies such as computer vision, hybrid AI and automated data labeling. Can you describe a specific project or application in which these technologies are used and how they shape the work at Continental?
As part of the publicly funded AI Knowledge project, we are researching hybrid AI methods that aim to combine conventional processes with AI processes and thus harmonize the best of both worlds. In the future, for example, route planning for highly automated and autonomous driving will not only incorporate observations from many thousands of real-world recordings in order to select the most intelligent driving direction using AI, but also the very conventional boundary conditions such as physical limitations of the vehicle and the environment, e.g. road width, road layout and braking distance.
You once emphasized that the speed of AI development challenges traditional methods of software development. Can you expand on this statement and tell us how you are tackling this challenge at Continental?
Not only the traditional, but even the modern methods of software development! These are primarily aimed at getting solutions that are already sufficiently established onto the market more quickly. While this approach may be sufficient for established solutions, with the large number of AI approaches available it is also important to select exactly the right and best one before the actual development begins - and this changes on an almost monthly basis. Efficient software development alone is therefore no longer enough. Rather, very good networking with science and the start-up scene is also necessary in order to keep pace with developments. With our AI Lab in Berlin, we are killing both birds with one stone, as described above, and optimally complementing our already well-established software teams.
How does the Berlin AI Lab contribute to Continental's efforts to shape the major trends in the automotive industry - automated driving, connectivity, electrification and shared mobility?
Shaping is a very fitting term in this context: we are involved in many of the most exciting of these topics through our research-heavy concept, which relies primarily on AI scientists and doctoral students. With our autonomous mobile robot, for example, we are mapping the entire chain: An electrified vehicle that, networked via a fleet management system, autonomously delivers parcels to a wide variety of recipients. Our employees here focus on the various aspects of such a system and thus form a strong team that is available to the corporation as highly qualified junior staff, e.g. after completing a doctorate. In addition to this focus, however, we at Continental make very broad use of AI technology, which, in addition to the obvious future topics, is also intended to simplify our internal processes, such as labor-intensive requirements management or testing, in order to make Continental an "AI Empowered Company".
Your lab is based at the AI Campus Berlin. What role do collaboration and exchange play on the campus? Can you give us an example of how collaboration has influenced your work in the AI lab?
Thanks to the many events and presentations that take place several times a week, it is very easy to get in touch with colleagues who work with similar technologies. For example, we regularly exchange ideas with neighboring companies in the Safe and Secure AI Community on these topics, which are so important for our industry, or regularly receive the latest input on current hot topics through the Paper Discussion Group. Without naming names, we have also already entered into discussions with other companies represented here who, thanks to their experience and specialization, provide services that we would not be able to provide internally.
Can you tell us about some of the projects or applications developed in the AI lab that you are particularly proud of? What can we expect from your lab in the near future?
I have already presented two above - a third deals with the question of what "good" data must look like for an AI system and which methods can be used to prepare and make it available in the best possible way. And of course, as is typical in the automotive industry, all of this should work in real time and preferably directly in the vehicle. In current development, this is one of the driving problems that we hope to solve, both in terms of development costs and the quality of the results. We have also recently won a "Best Industry Paper Award" in the field of explainable AI, which is quite remarkable for a lab that is only two years old. We expect to see the first major results as soon as the first of our employees complete their doctoral theses.
How do you promote a culture of innovation and entrepreneurship in the AI Lab and what impact does this attitude have on the way you work and the results of your team?
Through a very flexible work and task structure, as well as through our innovation process, we specifically give our employees the freedom and also the recognition when proposing and realizing ideas. This innovation process allows us to realize ideas as prototypes and then promote them in our business areas. At the same time, however, every idea has to undergo a tough selection process at the proposal stage, which also examines its commercial viability and whether the idea is in line with our Continental strategy for the future. When an idea is realized, the team is naturally many times more committed to the implementation than is usual in a company, which you can see from the enthusiasm in their eyes when they see the first results.
Thank you very much for the interesting interview!