Interview Transcript
Q1: Tell us your full name, your date of birth, and where you were born.
Birgit Vogel-Heuser: My name is Birgit Vogel-Heuser and I’m working at Technical University of Munich at the moment. I’m a full professor there. I’m from Germany, born in the area of Cologne—a city most of you will not know anyway.
Q2: What was your upbringing and childhood like in terms of your family background?
Birgit Vogel-Heuser: My parents were a cook and a butcher. I was raised as usual in Germany and took the normal pathway: higher education, and then I decided to go for electrical engineering.
Q3: Where did you go to school and do your PhD?
Birgit Vogel-Heuser: I did my PhD at the University of Aachen.
Q4: Did you work in industry after that?
Birgit Vogel-Heuser: Yes. I graduated as an electrical engineer, completed my doctorate in mechanical engineering, and nowadays I am more or less a software engineer.
After graduating, I went into industry for about ten years and worked at a company that does worldwide plant manufacturing. I had the honor of being the CTO. It was a mid-sized company with about 280 employees when I started and 450 when I left. So it was quite a challenging and growing period. I had to learn a lot and was responsible not only for technical issues, but also for details of the R&D department and worldwide startup and commissioning of plants.
Q5: What made you interested in engineering in the first place?
Birgit Vogel-Heuser: To be frank, it was my math teacher. He told me I should consider engineering. I went to career counseling to look at courses, and after a tenth-grade school class, I had the opportunity to do an internship at a CNC machine company. That was finally it. When I saw the CNC machines and how they are programmed, I thought: that’s what I want to do.
Q6: What was the first project you worked on in academia?
Birgit Vogel-Heuser: In academia, I had the pleasure as a PhD student—which is maybe a bit different in Germany than in other countries—to work on a very industry-related project. I had the opportunity to build welding robots for single workpieces together with a group of PhD students under other professors.
That was quite challenging. I built the user interaction part using a voice interface, designed the ergonomic human factors side, and handled programming alongside experts in welding and safety.
Q7: What was the main research question guiding that work?
Birgit Vogel-Heuser: The research question was simple in one way and very complex in another: How can we enable the use of welding robots—which are usually only applied in mass production—for very low lot sizes, say 1 to 5 pieces, in real industrial environments?
Q8: Moving between academia and industry, have you faced any major challenges in your career?
Birgit Vogel-Heuser: We all face challenges—a lot of different ones. I enjoyed being in industry; I could move things forward and have a lot of influence and impact. Now in academia, it is a different type of impact, but both have their pros and cons.
Q9: Can you recall any technology or breakthrough that changed the way you do research or engineering?
Birgit Vogel-Heuser: We are currently doing research on Large Language Models (LLMs). It is changing the way engineers work now and will work in the future.
However, there are nice studies describing why human knowledge and domain models are still necessary. We need to validate what an LLM provides—there are clear examples where LLMs just deliver false answers. It is a game changer, yes, but there have been many other small game changers over the past decades.
Q10: Who are the main funders of your research currently?
Birgit Vogel-Heuser: In Germany, we have a main agency: the DFG (German Research Foundation, similar to the NSF in the US), which funds fundamental studies. We also have industry-related projects funded by different government ministries, the European Commission, and direct industry funding. It is quite broad and has evolved over time.
Q11: How does having different sources of funding affect the way you do your research?
Birgit Vogel-Heuser: Let me answer that with an example. When I started as a young professor, a company called me saying, “We have an operational problem with a very sophisticated plant. Can you help us? We want to diagnose it like a doctor analyzing an X-ray.”
We started developing 3D visualizations of an ethylene reactor to identify hotspots. That began as a practical industry-funded application. From that project, we identified fundamental research questions, took them to the DFG, and got a fundamental research grant. After gaining those theoretical insights, we moved back to practical applications.
For me, research is a back-and-forth between fundamental and applied research. If you find requirements in real applications that you cannot address, you do fundamental research. When fundamental research needs validation, you involve industry through formats like DFG transfer projects. As automation researchers, we cannot just stay in theory—we have to apply solutions to real industrial problems to create true impact.
Q12: Could you share some highlights of your real-world applications?
Birgit Vogel-Heuser: One highlight was when we introduced object-oriented software engineering for future automation systems back in 2003. We helped deliver a tool that became commercially available around 2011. Many companies used it over its lifecycle until 2025, when it was retired. That shows the lifecycle of research ideas: a good idea gets adopted, serves its time, and eventually evolves as new trends emerge.
Q13: Has there been any surprising incident or unexpected outcome in your work?
Birgit Vogel-Heuser: Yes, many! As a professor, you pursue a research vision to ease automation or optimize plants. Sometimes you try for years and wonder why a solution isn’t accepted. It might be academically elegant, but industry rejects it due to practical constraints.
Another realization is that researchers are often too early. For example, my colleague Peter Göhner and I started working on multi-agent systems in 2003. Back then, people thought it was okay, but not essential. Now, over 20 years later, the time for multi-agent systems has finally arrived under new terminology. If you have a good idea, you shouldn’t give up just because industry isn’t ready for it yet; it might just take 10 to 20 years.
Q14: What advice would you give to students who aspire to reach a position like yours?
Birgit Vogel-Heuser: I never planned to be a professor or a CTO—it happened step-by-step.
First, work hard. Second, find the right mentor who will support you at critical junctions in your career. There will always be times when you feel like giving up. My advice is: never give up, trust yourself, and if one path doesn’t work, try another.
Q15: What do you consider your most important contribution to future automation and robotics?
Birgit Vogel-Heuser: I was recognized as an IEEE Fellow for contributions to intelligent, adaptive factory automation systems.
While many PhD projects focus on ideal, straightforward operational scenarios, my team loves to cope with faults because that is what happens in real industrial settings. My main contribution lies in developing software engineering methods and architectures that handle faults, providing the foundation for self-adaptive automation systems, including modern LLM-driven architectures.
Q16: Was there anything you would have liked to pursue, but had to set aside?
Birgit Vogel-Heuser: Many things have had to be reprioritized. For instance, we once had a PhD student who struggled for six years to secure funding for his topic. After eight years total, he managed to finish, but we had to reschedule and adjust our approach along the way.
Q17: Looking at automation and robotics over your career, how has the field changed?
Birgit Vogel-Heuser: It has changed dramatically, and yet in some ways, it hasn’t changed at all.
On one hand, we see massive changes with robots, cobots, and LLMs. On the other hand, complex tasks still require human manual labor, and industrial plants operate for 30 to 50 years. Whatever new technology we introduce today may take three decades to replace existing infrastructure. We still have legacy plants running, and we still face the reality of setting up industrial sites in remote locations—the middle of nowhere—where conditions are tough and require resilient people.
Q18: Where do you see the field going over the next 10 to 20 years?
Birgit Vogel-Heuser: It depends heavily on international cooperation and standardization. Currently, different continents are proposing different architectural standards.
For global plant manufacturers to join forces, we need cross-continental standards, which we lack today. Whether we overcome these hurdles depends on policy and global alignment on techniques and frameworks.
Q19: What core philosophy guides your approach to research projects?
Birgit Vogel-Heuser: Real fun and impact come when you step out of your comfort zone. From a systems engineering perspective, you shouldn’t just look at a single robot or CNC machine—you must look at how it integrates into the broader network.
That includes technical networks, business models, and societal impact. We need mechatronics, software engineering, and economics working together. Researchers should broaden their view to understand how their piece contributes to a prosperous society.
Q20: Having led projects in both industry (as a CTO) and academia, how do project dynamics compare between the two?
Birgit Vogel-Heuser: The nature of pressure is completely different.
In industry, as a CTO, you bear direct responsibility for the company’s financial survival and employees’ jobs for the next 5 to 10 years. You must succeed, maintain technological advantages, and make high-stakes calls under severe pressure. A wrong architectural choice can cause massive financial or operational trouble.
In academia, while there is still pressure, it is more flexible. You can choose which research questions to tackle and decline unpromising projects. If an academic project fails—for instance, if we underestimate constraints when trying to perform regression verification on complex industrial code—the worst outcome is an unaccepted paper or needing to pivot research. You don’t face the risk of 450 people losing their jobs or industrial safety accidents.
Q21: How does your practical background in industry affect how you choose academic problems?
Birgit Vogel-Heuser: Having a deep understanding of real industry requirements and constraints is a treasure I have used throughout my academic career. It builds trust when collaborating with industrial partners, allows me to ask practical research questions, and helps me move back and forth between fundamental concepts and industrial applications.
Q22: How do you incorporate Human-Robot Interaction (HRI) and human factors into your automation research?
Birgit Vogel-Heuser: I frequently collaborate with colleagues in Human Factors because combining domain expertise yields far richer solutions.
For example, we recently used human knowledge graphs to constrain Large Language Model predictions. In one case, pure data analytics produced an incorrect diagnosis, but human operator domain knowledge correctly identified the root cause—and we proved the humans were right.
Our goal should be to support humans rather than replace or overrule them. In applications like heavy construction machinery, reducing the degree of automation allows skilled operators to leverage their expertise while receiving decision support. Automation should assist, not discourage or overrule, human operators.