Interview with Heike Vallery

Heike Vallery: Well, yes, my full name is Heike Vallery. I was born in Duisburg, which is the city in Germany. Family life, well, my family was always very supportive. But there’s nothing in particular that influenced me. Well, my parents are actually both historians. And I’m the only one who went into anything engineering. So–

Q1: Do you remember what was the initial interest in robotics when you first heard of robotics? Maybe it could be a book or a piece of culture. But do you remember your first robot?

Heike Vallery: Hmm, my first– no, I don’t remember. I’m sorry.

Q1: When did you then decide that robotics was the career path for you?

Heike Vallery: Well, that actually happened towards the end of my master’s studies. So I had always liked mathematics. And I was looking for something where I could use mathematics in an applied way. And that’s why I chose mechanical engineering. And then actually, I started– I specialized in automotive engineering. And then, towards the end of these studies, I realized that I didn’t want to dedicate my life to building cars. I wanted to give back to society in a more meaningful way. At least I thought that I could do that better in another area. And then I actually traveled to Argentina for an internship and also to study there for one year in total. And there, I realized that people with any sort of impairment were much less visible in daily life. And well, I didn’t really– yeah, I had no real explanation for it. But I thought that it might be due to the infrastructure.

​ And then I thought about building something that would help people with impairments navigate also less accessible infrastructure. And then I actually asked at the university where I studied whether I could build something like an exoskeleton. And I wasn’t even aware that much of this was already happening. And they referred me to Munich. And that’s where I came into contact with more of that research. But I would say the most influential person was actually Dejan Popovic. I don’t know whether you know him. He was one of the Yugoslavian scholars working on assistive devices, neuroprosthesis. And I went to a summer school that he organized in Kotor in Montenegro. And that’s where I first really came into contact with the community of people building technological devices to assist people. And yeah, I would really say that he was probably the person that taught me the most and also influenced me the most in the choices I subsequently made.

2

Q1: Excellent. Thank you. You answered say subsequent question. And so your interest in robotics or your first robotics projects really begin with assistive robots. Are there other kind of central research agendas that you have had over your career?

Heike Vallery: I would say my agenda that has developed over the years is more methodological than an application-oriented one. So I realized after a while in the field that control can only fix so much. And that it’s super important to think about hardware and control together from the start. So to have a systems engineering view when you design a new system. And well, they would say that’s my first mantra is to always look at hardware and control together. And then the second one, which developed subsequently, is to always look for an absolutely minimalistic solution and to avoid wasting resources and also overburdening people with technology that’s just over-engineered. And always look for the simple solution that fixes the most relevant problem.

Q1: Yeah. I can see many reasons for that. Do you think there’s something about your training as a mathematician that encourages that emphasis on elegance in simplicity?

Heike Vallery: Well, I wasn’t really trained as a mathematician. I just liked maths at school and then studied mechanical engineering. So I’m really not a mathematician. But what I like, I definitely like analytical solutions. So I still enjoy the possibility of deriving equations of motion by hand and not relying on numerical simulations all the time. Because sometimes, actually, you can derive a close solution. And I really enjoy that. But that’s maybe more a pet project than it has a lot of relevance.

Q1: There have been tools that have emerged in the last several decades. I’m thinking about either technologies like Connect or conceptual models like LLMs that have been very significant in your work. Are there tools that are emerging now that you’ve heard of?

Heike Vallery: Well, I think you pretty much answered the question already. What’s emerging now, obviously, is all these data-driven methods that are a revolutionary effect right now. But in the past, I would say what, in my field, I recognized as particularly revolutionary was the emergence of very small and very cheap sensors. So just the ability to put IMUs on any moving objects at a negligible cost has made an enormous difference in what you can do. And I would say that also examples like technology that isn’t even necessarily a full robot can still be extremely advantageous to people. So for example, I spent almost 15 years building a gyroscopic backpack trying to help people balance. And well, we built this device. And it kind of helps people balance. But it’s still an enormously complex system. And at the same time, I also worked together with a company that was building airbags to just protect people’s hips from the impact. And all we needed there was an algorithm, just sensors and an algorithm and very simple hardware that would just inflate during the fall. And this type of technology is so much simpler, 3

so much more accessible. I wouldn’t even really call it a robot, although it might qualify as one. But it’s not the traditional system with actuators and sensors and some processing in between. But it has a tremendous effect of actually helping people.

Q1: Can you say a little more about your interactions with commercial companies and how that relates to other sources of funding? What has been most important for you?

Heike Vallery: Yes. Working together with companies, I wouldn’t see it that much as a source of funding, but as a way of getting technology to people and also building the right technology. Because companies tend to be a very good filter of ideas. If they don’t see a business case for it, then actually it might not be that likely that this will ever reach a person. Of course, it could also be just a matter of time. But in my experience, having a company actually invest in a product is a very good indicator of how valuable this technology actually is, and how likely it will actually reach somebody. So, of all the devices, of all the ideas that I ever had that reached the market, only those where a company was involved very early on actually made it. Yeah, I think I didn’t phrase that correctly. I think only those devices where a company was involved early on made it to the market. And I had many more ideas, but they got stuck along the way. And I think that’s just due to the fact that– well, actually two facts. Companies, A, they are good filters, and B, they guide the development in a very good way towards something feasible.

Q1: If I could expand on this idea of being stuck, as you think about things that you have wanted to accomplish, research and otherwise, what have been some of the challenges or obstacles? It could be the absence of technologies, or funding or institutional or political issues? Have there been significant obstacles in your research?

Heike Vallery: Well, obviously there’s always obstacles. And many ideas that I had didn’t make it, mostly because, lack of funding. But I wouldn’t really say that there was a systematic problem. Maybe a perceived problem is that I always found it easier to acquire funding for complex technologies compared to very simple ones. So for the very simple ones, reviewers didn’t believe that there was really funding necessary or that this was really an idea worth funding because it seemed so trivial. So actually, I had a couple of trivial ideas that just never made it. Although I think they could have been helpful, maybe they also were not placed in the academic system correctly. Maybe there should have been a different way of getting them out to industry. But wouldn’t say that I have met very big obstacles, mostly I’ve been really lucky in getting funding. Yeah.

Q1: Can you tell us a little bit about what you think your most significant contributions to the field have been?

4

Heike Vallery: Most significant contributions. Well, I think to the field of actuators, I think I have contributed quite a few small improvements to the idea of series elastic actuation. How to control them, what their limitations are, how to modify the mechanism to fit even better for certain applications. So I think that’s where I have made a modest contribution on the actuator side. And in terms of applications, I think my biggest contributions were the development of 3D body weight support systems. So those are systems that people can use to train after, for example, a stroke or a spinal cord injury. And I was part of actually two devices, or development of two devices, the FLOAT and the RYSEN. And they are both in the market now. And I think they are both used by individuals with motor disorders. And I think whenever I saw people using it, it gave me a huge satisfaction of actually having contributed to something meaningful. Yeah.

Q1: That actually brings me to my next question, which is about how you’ve seen your academic research move out into the world. That seems like an obvious way in which that has happened. Are there other aspects of this movement that you’re interested in, or proud of, or have been very surprising to you?

Heike Vallery: Well, I think maybe a bigger influence is why are the students that are educated over the years? I taught quite large bachelor’s courses, like at TU Delft, I taught dynamics. And I’ve very much enjoyed that. I wrote a book at some point together with a colleague of mine, Aaron Schwab. And I still actually sometimes meet students who took that course, even here at ICRA. When I walked into the doors, I saw one of the volunteers actually walking and taking my badge. And I wondered, how would he know my name? And he said, like, a couple of years ago at TU Delft, I took your course. Apparently, at least he still remembered me. Unfortunately, I didn’t remember him. But I think that maybe there I’m even– I have more impact than with my academic work. But yeah, difficult to say.

Q1: That must be very rewarding.

Heike Vallery: Yes.

Q1: As you think back on the field, as you’ve experienced it, and look forward to the near future of the field, what do you think has changed most thus far? And I’ll follow up then and ask you what you think is likely to change.

Heike Vallery: What has changed most? Well, I would say, we are almost in a sort of Gartner Hype Cycle, but we have passed the peak of inflated expectations for many of the technologies. And we have passed, well, maybe multiple peaks already, quite a while ago. So when I started in this field, people actually believed that rehabilitation robotics was going to be something that made people recover better or faster. And that was actually disproved by a couple of larger studies that were made with very high expectations, but that showed that conventional therapy is 5

just as good or sometimes even better than rehabilitation robotics. So then we reoriented towards not promoting these robots as therapist replacements or better therapists, but towards them being tools for the individual’s training, and for the therapist’s coaching. And I would say that was a major paradigm shift.

​ Recently, also, I would say there’s been a trend in the exoskeleton community, moving away from therapeutic wearable exoskeletons towards more devices that can be used in the workforce, for example, to support people with heavy loads. I see that a little bit skeptical because I’m always afraid that there might be effects that we’re overlooking right now, let’s say negative effects, when you start introducing such technology for healthy people to augment their capabilities. I don’t want to speculate too much about it, but I’m just afraid that there might be damage to the biomechanical function or the musculoskeletal system that we are not currently foreseeing from long-term use of these systems. But generally, I think it is a clear sign that the field is maturing and there’s commercial interest. And I also see that the designs are much more elegant now, much more simplistic, much more targeted, much more thought through than they used to be.

Q1: You alluded to this already, but if I could ask you directly to think about the influence of recent developments in generative AI on the future of the field, what is your take on that?

Heike Vallery: Well, obviously, there has been an enormous development in the digital world so far. So there’s an enormous AI boom. But it has not yet made the step into the physical world. So I think robotics can be a vehicle for that. But for that, we might even have to build robots in a different way. Because we have built them keeping in mind which type of technology we have to control them. Like, we build robots thinking about rigid structures, or at least that’s still a frequent paradigm of building robots. And then we think about modeling them with differential equations. But once we embrace AI, we don’t really need to model them anymore with these engineering tools. So we might also have much more freedom in how we design robots. And obviously, people are already building all sorts of squishy and soft systems that will be much more easy to handle once you let loose, let go of model-based approaches.

Q1: Since you work with augmentation technologies, I’m curious, do you work in your labs with human subjects? Does somebody else do that work? And how do you think about that relationship between the human and INAUDIBLE?

Heike Vallery: Yes, I personally like to stay on the side of developing technology and always collaborate with people who are more skilled in doing studies. Sometimes I do smaller studies also in my lab, but it’s not my strength. So I mostly work together with movement scientists or with clinicians, therapists who have all this knowledge and also want to take the leading role on actually doing studies. And I think that’s critical to have a very strong background in doing these studies, which I don’t think I have.

6

Q1: You mentioned several times your interactions with students and the way that you’ve influenced the world through students. What kinds of preparation do you think students coming into the field today would need to most prepare them for success?

Heike Vallery: Well, analytical thinking will still remain an important asset. Flexibility, the ability to learn new things quickly, the ability to communicate and integrate. Well, I hope to think that still mathematics will be very helpful. But I don’t want to make predictions of what people really need to know in 10 years. I think the more flexible they are and the more able to acquire knowledge and skills quickly, the better prepared they are.

Q1: Are there research problems that you’re interested in taking on, that you have not yet been able to pursue, but would like to pursue in the next phase of your career?

Heike Vallery: Well, I sometimes think that I would like to have the opportunity to start again. And I don’t think that’s really possible. But I did now take over an institute that is not really so closely related to what I did in the past. So I now had the control engineering institute in Aachen. And that – sorry, I even said it wrong. I had the Institute of Automatic Control, which does control engineering. And robotics is just then, just one aspect of this. And also the institute is much more oriented towards algorithm development and much less towards building entire systems. That’s normally only done in collaboration with others.

​ So I am, in a sense, already making a move to do something different. And what I like about it is that I can now start getting into other areas that I think are really relevant for today’s society, especially the climate crisis. And so, for example, we look into energy, sustainable energy and control aspects of that, like, for example, wind turbines, how to control them, how to make the grid more resilient, to the fact that there will be a lot more of these renewable energy sources in it. And this is very far from my own education and background. So I’m now learning a lot more about these new applications, which is not so much robotics anymore, but I think still a way of contributing to society with the skills that I have and, well, particularly enabling the people in my institute to to contribute in that way.

Q1: You discussed a wide variety of interests and problems and methodologies. I’m curious what for you, describes robotics? So robotics as an umbrella term, but what is the kind of essential aspect of a project that, that makes it a robotic project?

Heike Vallery: Mm hmm. Well, interestingly, I recently looked up some definitions of robotics and they don’t. There’s some norms, of course, the ISO norm. But but they don’t all agree. What, what you at least see in common is that it’s a system that interacts with the environment and mostly in order to to perform a certain task. So it’s somewhat goal oriented. And normally you also see it in the definition that it has sensors and actuators, and some some way of reasoning and 7

connecting the sensory experience, and connecting it to to an actuator response. And I would say, in my own definition, that is also a core part of robotics, the fact that it can sense its environment and influence its environment in a physical way. I would deviate a little bit from the definition in saying that the goal or task has to be explicitly defined. So I also would consider devices, robots that have sort of emerging behaviors that are interesting to observe, but not necessarily, from the start, goal directed. But maybe in that way, I deviate a little from the from the main definition.

Q1: Well, that is interesting. I have also been looking at definitions lately and have found similar ambiguity. As we are nearing the end of our our time together, are there questions, or topics, or things that you would like to say that we haven’t yet talked about?

Heike Vallery: Well, I think that the field of robotics is very much gaining in importance and should be aware also of the environmental impact that it will have. And I find that is currently yeah, this gets too little attention. Well, we know about all the effects of resources that are already used for smartphones, which are a magnitude less complex devices than robotic devices. And if we start building now more and more of these, and spreading them, and particularly spread them for purposes that are more to improve convenience of life, but not critically necessary, then I think we have to ask ourselves ethical questions about, well, use of resources and use of energy for this, for this field.

Q1: Do you think that the field is well prepared, either kind of as individuals or as a set of institutions, to engage with these ethical questions? And if not, what would you like to see in the field?

Heike Vallery: I think, well, there are ethists, ethicists. How do you say that? Ethicists? Yeah, yeah. Well, there are ethicists asking themselves questions about robots and engaging with roboticists. But I see them more as a side note often. And I think, at these types of events, maybe not that these events, but in the research people do, I see the ethical questions too little ingrained or part of the whole design process, which is not out of malice, but just because it hasn’t been such a central or urgent question so far. So, so I think that we could we could benefit more from from inviting these people who have very valid thoughts and concerns more into our community. But maybe I’m also just not attentive enough to notice it. That’s also a possibility.

Q1: Well, thank you. We are at the end of our time. Gerald, is that technically OK? Thank you. 8