Interview with Bram Vanderborght

Interviewer: To begin with, would you mind sharing some background details? So your name, your date of birth, where you’re from, anything about, you know, your family, where you came from, early schools, things like that.

Bram Vanderborght: Yeah, I’m Bram Vanderborght, I’m a professor at the University of Brussels, and I’m affiliated with Eilmik Large Research Institute on Microelectronics. And I come from parents who did chemistry, so my grandfather was teaching in technical school. So he was very happy that his grandchildren were all working in technical field again, and I also learned a lot from him about technical stuff. But also, from my parents, we also did a lot of hobby projects, and so that’s how I get interested in technical and engineering.

Interviewer: What was your early educational experience like? So high school, college, any information you’d like to share about that?

Bram Vanderborght: Yeah, so as a child, I was a lot interested in things that move, and so I started, I wanted to build model airplanes, but I was not allowed for my model. So the first thing I was building was a model boat that really worked, and so with my father, I learned a lot about electronics and electricity and so on. But then I couldn’t hold anymore, and then I started to build model planes, first from existing kits, but then I also started to design my own kits, even like a glider with solar panels. But yeah, that’s a very long time ago that I was experimenting with that, and then yeah, I really wanted to become an engineer. So I went to the university to do mechanical engineering, and then I had the opportunity to work on my master thesis on a big bipedal walking robot with artificial muscles instead of electric motors. And although I never thought of an academic career, then I started to do a PhD. I went to Cuba in Japan to work on a humanoid HRP2, and then I finished my PhD, and then I did my postdoc at the Italian Institute of Technology, and I had the opportunity to become a professor at the University of Brussels, and so then I became a professor.

Interviewer: Would you like to share anything about any mentors or colleagues along the way?

Bram Vanderborght: Yeah, so my supervisor was Dirk Lefebvre, and he inspired me a lot about the mechanical aspect, so he was always a deep dive in the mechanics. And so that we also continued during our research to focus on hardware, and so now it’s really best worked AI and foundational models and so on, which I think are very important evolutions. But I think they can only work when we have reliable hardware under it, which I started as a mechanical engineer, being mechanics and especially actuators. But now the work extends also to other core robotic technologies for a robot, which is not only activators and variable impedance activators and locking mechanisms, and so on, but also new materials, new sensors. And because I’m now in IMEC, also thinking of what will be the chips of the future for humanoids.

Interviewer: So very much a hardware-focused scholar.

Bram Vanderborght: Yeah.

Interviewer: As far as work experience you’ve had, so you described the arc you took through the academic field, do you want to say anything about where or with whom you’ve worked, when or what departments?

Bram Vanderborght: Yeah, so I did in the mechanical engineering with supervisor Dirk LeFaber, but then I did also with Professor Yokohisa in Sukoba, in AIST, and so there they had the humanoids. And of course, when I was doing my PhD, I had a bipedal working robot, and then he had also Asimo Komi, one of the first really feasible humanoids. But of course, the research version of that was HRP2 in Japan, and it was really a dream to work on that very expensive humanoid, but it’s very impressive that humanoid is still used for decades after it. So it’s really an enormous, reliable platform and also an inspiration for many work and all the humanoids that you see over there. And I had the opportunity to work for Darwin Caldwell, because at that time it was the Italian Institute of Technology was brand new. It was built in the mountains in Genova. And yeah, at that time we still had a lot of place, a lot of desk space, and now it’s crowded with a lot of persons. It really grew amazingly and also a lot of important robots from now, from IIT, started at that time. So it was a wonderful time to start from almost nothing and help Darwin to build up a new lab in Italy. And then I had the opportunity to go back to Brussels because a professor retired, and then I could, together with Dirk Lefebvre, who now is retired already a while, but still active in our spinoff company, had to build a new lab. And yet there I focused a lot on multidisciplinarity.

Interviewer: Have you already mentioned what that spinoff company is?

Bram Vanderborght: Yeah, we have several spinoff companies in development, but maybe we can come back later.

Interviewer: Sure, absolutely. So you talked a little bit about how you got interested in robotics through engineering and an early interest in flight. Seems that move. Along those lines, were there any kind of events, books, films, media that also inspired you or sent you down that track in the first place?

Bram Vanderborght: Yeah, I went as a child, my parents for a while moved to Denmark and there is a famous Lego lens. And apparently I was completely paralyzed and overwhelmed. I couldn’t process like that dream world. And so later we visit again. And I still have a picture where I’m like passionately looking to Lego robots arm. And also I still remember going to Disney. And I don’t know the name of the robot anymore, but above the crowd to enter a certain attraction, there was also it was probably more animatronic than a robot. But the fluentless and the way it could captivate the audience was also very inspirational. So, yeah, when I when I wouldn’t have been in robotics, I would be probably in aviation and so on. But when I had that opportunity to do the master thesis, then I started to make the decision. I think I can have more impact in robotics than in aviation. And so I saw a lot of opportunities in robotics to to improve the world. We have a lot of challenges in our society, economy and planet where robotics could be part of the solution. And then I decided, yeah, I want to become part of the robotics community. And yeah, that’s I’ve been in everything.

Interviewer: So you described the sort of first robot you’ve ever seen, so to speak, if it was an animatronic. Would you count that as the first robot you ever heard of? Or would there be a different example of a more classically defined robot that that was the first one you ever heard of?

Bram Vanderborght: Yeah, that I don’t remember, honestly, anymore.

Interviewer: What was the first robotics project that you worked on? Can you share a little more about that? I know you’ve already mentioned a little bit about it.

Bram Vanderborght: Yeah, so the predecessor of my supervisor ended up in a wheelchair. And so it was quite sudden. And then we thought how we could help people to walk again. And so at that time, it was still the idea of to make very rigid robots. And so before my time, we had a hopping robot in the lab inspired by also the work of Mark Rebert. And it was called Ollie from One Leg is Enough. And when it hopped, it really apparently because it was before my time, the building shape because of the impacts. And then we started thinking before my time how we can have elasticity compliance in it. And then they started to have the development of the pre-nomatic artificial muscle, which consisted not of a netting with a rubber tube, because that has a lot of hysteresis. But with artificial muscles with pleats that unfold. And so they could generate very high forces at very low pressures and have that elasticity. And so when I came in, they had the idea to make a bipedal walking robot with it. So during my master thesis, I integrated together with two other researchers, this is in a bipedal walking robot that we completely built from scratch. And the lab didn’t have a lot of money. So we really had to make everything ourselves. I was remembered that we made our own data acquisition cards. So I don’t know how I made it, but we made the card for the computer. And so we could read out all the data and we went to buy in the Ollie a cheap balance for the weight of a person to take out the string gauges and those kind of things to assemble that robot. And so it was a walking robot walking in a plane, because then we could use 12 muscles to have six degrees of freedom activated. And so we could then experiment how we can change stiffness to reduce energy consumption. So I tested that on a pendulum set up. And yeah, we then we managed to make the robot walk. And so I think it’s still a panomatic biped that’s walked. That was, I think, quite impressive still at that time. But then there was no real funding anymore for humanoids at the time we have changed in between. And so then we diverted more to again that dream how we can let persons walk again. And from there, the track on prosthesis, a bionic feet started not with problematic muscles anymore. But the colleague invented the Macapa activator where very simple. The stiffness of the joint can be changed. I like which change the stiffness in our joints. And we also started to develop exoskeletons. So we’re all robots. And so in order to further improve energy efficiency, we started to include locking mechanisms. So energy could be locked. We started to use parallel elasticity and so on. And from those two tracks, then one track was the development of a company, Accelus Bionics, that now has on the market bionic prosthesis. Now they have it for the foot. But now we’re transferring the IP for the knee. And a separate track was exoskeletons, where we developed exoskeletons for rehabilitation, but also occupational exoskeletons. So to support workers to work overhead or to carry heavy loads. Where I did the ideas how much of the energy we can save by using mechanical elements, like springs and locking mechanisms, and at the minimal amount of motors in order to provide extra energy. And so also that track is now in a way to create a venture called Husk in order to not only have very nice papers, but also create real impact on a society by having a company.

Interviewer: What was the time period when you were developing that first robot and going to Aldi, roughly speaking? And then when would you say was it that you noticed this decline in an interest in humanoid robots?

Bram Vanderborght: So I started my PhD, my mouse thesis in 2002. And so in 2003, I started my PhD. And so then at a certain time, Asimo made a running robot. I made a running robot, Asimo, that could run. And it was really like as a PhD student, what I’m doing, what is my role in this research. But of course, it was a very different approach because they all had to stiff activators and they had to use a lot of energy to absorb the shocks and to have the next jump. So that’s why we went to it to having passive elements that can store and release energy like our muscle tendon system does. So but we, of course, we learned a lot from that. And yeah, that’s always a bit research is going to I say it often to new PhD students realize what you start with. You’re going from problem to problem. And of course, also today we had a project that was rejected. And my wife said it’s impressive how you can deal with all of those rejected projects. But on the other hand, when you don’t shoot, you don’t also have accepted projects. And yeah, the impact that it can generate so that drives us forward.

Interviewer: Well, that is a good segue into a couple of related questions. You’ve already touched on this to some extent. But would you mind sharing a little more about the main research questions that you’ve been interested in how they’ve changed over time?

Bram Vanderborght: Yes. So our main research is how we can create impact for a better world. And so for us, the sustainable development goals of the UN are very important, which I see also on the three pillars. So it’s not only planets, but also society and economy. And of course, there are contradictions between them, but how we can balance them and provide services to the society with it. And so how we can create that in a robotics context. And so what for us is very important is how we can approach it from a multidisciplinary approach. And because a robot is an integration of many of the best technologies on material sensors, actuators, processing and so on. But since we often have to deal with humans, also the human social and medical sciences are very important. And so that’s why we started to gather all the different research group at the university, not only the technical, but also the social, medical and human sciences. And that creates robotics, which I currently heading. And so that means when there is a question of a company or we have an idea that we try to make multidisciplinary teams around it. So our robotics group is certainly not one of the largest, far from. But I think what’s unique is that we do it in a very multidisciplinary approach and that for different applications, which are, of course, also growing. And what for me is very important is, again, what are the hardware innovations we can do with a layer of AI and control in order to bring that to the society. And so that’s why I’m a mechanical engineer. So, again, I like to build things and to invent new things. And I think that also has an advantage from an economic point of view, because as mentioned, we are starting several startups on that. We can have IP and patents, which are, of course, very valuable to build a company and attract investors from it.

Interviewer: Now, you mentioned the three pillars and the SDG goals. You’ve also mentioned something about efficiency as an interest in developing the robots that you’ve been interested in. Does energy play any part in this sense of sustainable development or am I off base here?

Bram Vanderborght: Yeah, so energy efficiency is very important because all those robots consume quite a lot of energy. And I think we need to reduce them, especially also because a lot of robots have batteries and we need to avoid using too much energy because all those energy you need to carry in batteries, batteries which add extra weight. And so that’s why, for example, in our RedWorld robots, prosthesis, exoskeletons and so on, we focus a lot on how we can improve the energy efficiency. And so another technology from our lab, also inspired by Dirk Lefebvre, is the development of new gearboxes. And because motors spin at a very high speed, but what we need to do in robots are low speeds, but high torques. And it’s incredible how much torque a human body can develop. For example, when we do a small jump, it’s a torque of a Formula One motor. Of course, we don’t reach those speeds, but it’s enormous torque machines that humans have. But that’s the problem. Electric motors are not good in torque. So that’s why you need a reducer, a gearbox, and a lot of energy is lost over the gearbox. So that means if I want to hold a certain object in my hand, the motors, of course, serve to take that object. But of course, there is a motor here in the palm of the hand and in the elbow in the shoulder that becomes bigger and bigger because most of the torque is in the hand. Torques it needs to develop is to carry its own arm. And so if we can improve a lot of energy efficiency, then you can use much smaller motors and then you don’t need to have the next motor so big. So also there is a reduction possible, but then you don’t need so much energy. So you can use smaller motors, smaller batteries. So that means smaller motors again. And like that, you have a lot of improvement in the efficiency, but also in the safety of the robot because those elastic elements, of course, also improve safety of the robot. And like our elasticity allows to have impacts which don’t harm each other.

Interviewer: So you mentioned already some of the advantages of approaching things in a multidisciplinary way. I imagine that can also multidisciplinarity can present its own challenges. And I was interested in whether related to that or not. Any big challenges that you faced in your career so far, you know, they could be technical, institutional, social, any events or circumstances that have affected your research?

Bram Vanderborght: Yeah, indeed. I think trying to work over different disciplines is at the beginning very difficult. And so we also experience that a lot because we just speak other languages. And so just understanding each other at the start of a project is a challenge. Moreover, those different disciplines have also in their research field different needs. And so you need to tune the needs of those persons and the habits they do research to each other and understand each other better. And typically, that goes pretty slow at the start, can lead to frustrations, especially for PhD students who need to maybe investigate an exoskeleton or a prosthesis while the prosthesis is not yet available or fails and so on. But I think after a while, when you build up trust between each other and learn to collaborate and do it really with a long term vision to collaborate. Yeah, I think at the interfaces of different fields there, you can really make the breakthroughs. And for example, I explained my parents and uncle levels with all in chemistry. So during one of the Christmas parties, my uncle was explaining cell of healing coatings for metals he was developing. And then I said, I want that in robotics because, yeah, we are developing soft robots and, yeah, they have higher chances to be damaged. And nature has a very interesting approach. We are not over-dimensions. When we have an excess in force or a cut, yeah, we have often the capability to heal our cell of. And so I wanted to have that in robotics. And so then we started to work together with the material scientists. But of course, they were only able to produce at the start grams of the materials. And so we also developed that long term vision, like how we calibrate and first use what they developed. And then learn, develop the robotic demonstrator from it and learn, okay, how the material should improve, what are the opportunities, what are the challenges. And so that team gradually grew. And so now 20 persons are working on that dream of making cell of healing robots and cell of healing materials. So those materials are also sustainable because we can extend the lifetime of a robot part because it’s damaged. You don’t need an expensive repair or you throw it away. Moreover, those materials are typically network polymers. You can give very interesting mechanical properties like the tire of a car or back a lid, which is a very stiff material, where the old sockets and phones were made of. But you can never recycle them. So that’s a huge environmental issue that a lot of plastics cannot be recycled or not back to the original function of that material. You can use it in sound tools and so on. But our network polymer can be fully recycled. Moreover, we are using also biosources and studying biodegradation. And so also from that we have from an initial idea during a Christmas party with someone from a very different field. And we have now we’re now creating a spinoff company called Falance Technology that will bring this to the market to make, for example, cell of healing and cell of closing suction cups to also reduce the energy consumption of the suction cups, which are a lot used in industry and use a lot of vacuum power. So a lot of electricity to do. They can be damaged, but we can heal that. So this is a technology that also has spillovers in many other fields. Like, for example, now we’re also making cell of healing tires for bikes and cars. I often go by bike to work and it’s a frustration how many times I had a puncture and then in the first had to walk out of it. But now we have a cell of healing tires that we also will bring further to the market, which also has the sustainability reduce the waste of tires, but also even safety of bikers of cars because even persons have accidents, but also having sometimes even rape because they had a problem with their tire.

Interviewer: As a biker, I’m interested in this question myself. Would this be related to things like tubeless tires or is this an entirely different realm of materials?

Bram Vanderborght: No. So the idea is that we still use tires with air because they have superior properties and indeed several attempts to have airless tires and so on. But the properties of a tire with air are still unmet. But the problem is the punctures. And so we put a cell of healing layer inside the tire that when you have a nail, you take it out, the material heals and you wouldn’t even have noticed that a puncture happened.

Interviewer: Well, you talked a little bit about sort of serendipitous events that have moved research along. Have there been any tools that have appeared over the course of your work that have had a major impact on the directions you’ve taken or possibilities that they opened up?

Bram Vanderborght: Yeah, I think it’s important to investigate what are the core problems in a robot. And I was always, always interested how I can solve them by starting to look to opportunities in other disciplines. And so that’s also the reason why I joined IMEC. IMEC is a very big research lab on chips and it’s a chip lab of the world. And of course, we use a lot of chips to process information. So now it’s just only starting. I’m trying to persuade how we can develop dedicated chips for robots or humanoids. It’s a very expensive development. But typically, robotics, I find the field where we borrow or we take developments from other disciplines. For example, the Kinect camera had a huge impact, better and video processing cards and so on. And we think, oh, cool, we can use that in our robots. But we should also more and more push other disciplines to develop technology dedicated for robotics. And one of those things is that also, for example, touch sensors, for example, are very important. So that’s why we work together with Melexis, which is a company in Belgium that builds sensors for cars. And now they’re also, of course, wanting to enter robotics and they’re developing a very interesting touch sensor. And what I think is important is we want humanoids that are affordable. And those humanoids of 400,000 euros or 1 million euros are nice for research. But we want affordable robots that can work in our household or in our industries, in our offices. But that means that all the components from which it’s built from, they need to be super cheap, but also very robust. And that’s what’s happening in a car. And so I find it very interesting that such a company wants to enter the robotics field because I think they can bring a lot of these expertise. And so by working together with those companies, we can do that. But that’s also what I want to do with the chips. So now in Europe, iMac is developing chiplets because the development of a chip is very expensive. But if you can break it down in smaller modular components, which is then called the chiplets, then the development cost is cheaper. And depending on the certain needs of a robot, you can assemble with a modular building blocks a chip with some those chiplets. And so iMac is now developing it very started to develop it for self-driving cars. But I think there is a lot of similar needs for a humanoid or a robot with self-driving cars, like we need to interact with our car, like we need to interact with our humanoid. We need to do path planning to process a lot of cameras, for example. Of course, the humanoid has extra requirements, especially I think the need for touch, which I think is still underrepresented in those foundational models. So we need to have other sensors like touch. And so how we can develop the dedicated chips, which have chips dedicated for the algorithms to process this information and do the control of it.

Interviewer: Don’t worry about it. We can’t hear it. If a distraction needs stop, we can’t hear it or report it. Okay, good. It occurs to me that the line between robots and cars is increasingly blurry in a sense. You know, as you’re describing this this meeting of materials and meeting of techniques, it really is becoming a little unclear where you can draw that line, depending on the model of the car you’re talking about, I guess. Have you ever had any thoughts, and this is not an important, we can move on from this, but about the boundaries or the definitions of what a robot is or is not?

Bram Vanderborght: Yeah, for me, what I often have, I explain it often to children is the analogy with the human body or the animal body. We need to perceive the environment for that we need sensors, we need processors to take decisions, and then we need actuators like our muscles to influence that environment. And we need a physical body in the real world. So it’s not only a virtual thing in a simulation. So we need a lot of materials for this. And the form and the shape and the size for me doesn’t matter. And I think it’s like the richness of the animal world. We also have the richness in the robotics world where you have a lot of different forms and embodiments of robots. And so for me, even also self-driving cars, even if they’re not level five self-driving capability, for me, it’s indeed also a robot. It’s like when I don’t do it, but in my village, we have a lot of horses. When you drive a horse, we also give commands, but it’s still the horse that decides where to put the feet. So for me, also a car that has a lot of style of driving functions, although not fully self-driving, is for me also a robot. So for me, that’s a rather big scope. But I think the underlying core components can all be very similar. And how you develop that for certain applications, of course, that’s for the integrators that make a robot with those technologies.

Interviewer: Thank you. Moving on to talk a little bit about funding, if you wouldn’t mind describing where your funding has come from over the years, how you developed those relationships and how that funding has or has not influenced the work that you’ve done.

Bram Vanderborght: Yeah, that of course has a big influence. Often the government is regarded as something or someone that objects innovation. I’m not in that same line. I think the government is the biggest risk investor for technology, even the iPhone. All the basic technologies are originated from government funded projects. Of course, you need a company like Apple with a vision of Steve Jobs to make it something cool, as, for example, an iPhone. And so there, of course, is an important relation between government funded projects and investors, entrepreneurs, to bring that to the market. And so this is also the way we work. So we have we are very dependent on government funded projects, both in our regional, flat, Flemish level and Brussels and the European Union. So the large amount comes from that. And the challenges in our lab, how we go from very deep tech technology, very fundamental, newer ideas and how we can fund that sufficiently up in the system. Up in the technology readiness level so that investors are willing to take it up and bring it to a company. And there I see also a difference that in Europe, they don’t want to take those risks. And I find that the government has to have a lot of developments until I find very high tier levels before investors are willing to take it up. And I think there is a difference with, I think, Asia, but also the US that much earlier they take to take those risks and they have to have companies that for many years have losses, but want to keep that same direction. And so that also, of course, influences sometimes how we do spin off companies. For example, for our exoskeletons, we started to develop centralized interfaces because we need the intention to move because not only the postures, the motions are important because if I want to lift a heavy table, I cannot arrive it from only motions. I need to go to the origin of the motion, which is my brain activity, but it’s still very difficult to measure. So we go to muscle activities and interaction forces and so on. And so we meet with that centralized interface. We connected to cobalt. So we developed a rehabilitation device. So if the patient was doing some effort, we could assist it because if the robot moves without effort, you don’t have therapy. And then we were finalist of the KUKA Innovation Award and we wanted to make a spin off company. But it’s again a medical certified product like our Bionic Prosthesis Company, a long investment pass, uncertainty on certification, uncertainty on the repayment process also in Belgium of physiotherapies. And so that’s why we pivoted now to a company called Schenetics that makes sensitized sports garments to see for elite sporters the risk for an injury. And so that’s also how we moved from robotics to another field where there is quicker time to market and less risk for investors to have a return on their investment. And so of course, like that, it also influences, yeah, also certain parts how to go to a spin off company.

Interviewer: Have any of those dynamics like funding structures changed much over the course of your career, you know, like proportions of money coming from government or private entities or anything along those lines?

Bram Vanderborght: Yeah, so in the beginning of the lab, we almost had complete dependency of government funded projects over the years that changed with more also income from companies and that companies want to work with us in order to develop certain technologies in the value chain, both from technology providers. For example, not only we work with me like this resource of voxel sensors, which is a company working on ultra low latency and ultra low latency 3D cameras. So they have a very interesting technology that they only make a laser dots doing this as you features in the environment. So it’s only one dot of energy and not throwing a lot of energy in the environment. They have the feedback, but it goes very fast. And so that’s why it’s ultra low latency, because if there is an obstacle, for example, for the drone, you don’t need to have a complete representation of the scene in order that the drone can already divert from that obstacle. And so also with that company, we have now collaborations in order to not only have the technology, but also the SLAM algorithms and the control in order to have the pipeline from the sensor to the activation of a drone or another robot environment.

Interviewer: Is there anything more you’d like to talk about? You’ve named a number of different collaborations or partnerships or spinoffs with companies. Is there anything you’d like to add to what you already said along those lines?

Bram Vanderborght: Yeah, what I think is very important, what we changed over time. In the beginning, we did a lot of research. And then we thought, we can make maybe a company out of it. And then started to do voice of the customers and finding investors. And then often it turned out, yeah, we maybe burnt money on research that is not in interest for the customer or the investor. And so due to this experience, we tried to do it now much earlier. So, for example, robots are very slow. And that’s not only for human operators, very frustrating, but you also kill the return on investment because robot moves so slow. Yeah. For example, we did a case for voluxwahr, for a certain piece. And we had a screen to tell the operator what to do. And it worked all very well. But it couldn’t keep up with the speed of the factory, how cars are built. And so we saw that in several of our projects. So the PhD researcher started with doing 30 voice of the customers with different companies, big companies, end users, robot manufacturers to understand what is the problem that we feel. And so that’s why now we’re developing also a spinoff company, Savebot, which is a perceptoring consisting of different sensors. So independent on the environmental conditions like fog or rain, we can have an occupancy map of what’s happening around the robots. So that robot can work full speed. And when someone comes closer, the robot will slow down and stop or take another route. And so we learned that much earlier, we need to take the customers or the future customers on board, but also the investors. So from a very early stage, we do friendly pitches with investors to see how we further develop the technology, what we need to protect, what we need to keep secret, what we need to do open source in order to have the highest chances that this spinoff is successful. And not burn a lot of resources to develop a project and the investors say, no, you need to do that. And so because we have several spinoff developments, it also inspires other researchers to see how it’s possible. And now they also come up with own ideas to develop their company, which I think is very nice to see.

Interviewer: So would you say the it sounds like you’ve described differences between the direction that state funding sends you in and investor funding can tend to lead you in because obviously they’re they’re sort of imperatives or incentives are different. Would you say that the emergence of more investor led funding has had a significant effect on sort of design choices or directions that you go in or not so much?

Bram Vanderborght: Yeah, so, of course, it’s always a challenge. Those investors will not invest the very low levels. So my continuous job is always to write grants after grants in order to have our team funded over the whole level. And also you need big sizes of teams have these different disciplines combined. But I think now there is indeed it’s good that there is increased interest from investors, I think, in robotics due to the rise of artificial intelligence that went from creating text for anything. The large language models than images and videos. And the next wave is, I think, physical AI. I think investors starts to realize that only investing in the AI companies as, of course, a false time to market and a team can develop cool things. But maybe chat chip material is a new version where that function is automatically and for free. And what is the IP of the company? So I have now the feeling that investors starts to understand that investing in hardware and companies with IP has also a lot of value. And I think it’s again a combination of what is hardware, where is the IP of the company? And of course, an AI lead around it to make it functional in different environments and so on. And that I think is a very strong combination that robotics has. And I think that’s important towards the future for investors and future researchers.

Interviewer: I just want to take one quick pause. I don’t know when you need to be out. I had heard both half an hour and we’ve done hour long interviews before. I just want to make sure we’re not going overtime for you.

Bram Vanderborght: No, no, it’s okay. I booked an hour.

Interviewer: Okay, great. Terrific. Good to know. I’ll make sure we’re out of your ad. There’s only a few more questions. So we should be good. So I guess along the lines of directions of research and funding and the sort of paths taken, are there any paths that you haven’t taken any kind of things that you would have liked to pursue but didn’t get the chance to?

Bram Vanderborght: Yeah, one of the things is I started with activators and I had a prestigious grant, which is an ERC starting grant. And there the idea I had was now we have one motor per joint. But if you see to a muscle, it consists of a lot of muscle fibers. And so when you see, for example, the energy for a chip, then the basic component is a transistor. And you add a lot of transistors for a very powerful chip and a few for less powerful chip. The same for a display. You have a LED and you have a lot. And then you make a display. But for a robot to activate, you need a whole catalog of big and small motors to assemble a humanoid. While the basic activation technology in nature is a muscle fiber from a fly almost to an elephant. And depending where you need to activate, you add more and more of those muscle fibers. So then I had the idea I want such a transistor for activation where we have not one motor per joint, but we have a redundant amount of motors per joint. And so like that we can also improve the energy efficiency because the energy efficiency of an electric motor is very good. But in a very small region at high torques or high speeds. Yeah, the energy efficiency drops a lot. And for our robots, we continuously use the whole working area off of the of the electric motor. And so you have the dual motor like for example in hybrid cars, where you have a combustion engine and electric engine. But also now there are bikes where you have the dual motor where you spread out the working point of the motor so that the overall energy efficiency is improved. And I wanted it to be even even like 10 or 100 or 1000s motors per joint, which I did in artificial lab settings, have several motors, but I never reached the capability because the manufacturing issues to make that didn’t evolve or were not available to really make such muscle fibers and like that combined. And I still a bit my dream once I still have a time before I retire and I had to maybe realize that another crazy idea is maybe to make robots that you don’t need to recycle but that’s like a human body or an animal can be composted. We use now a lot of earth materials to build our robots with neodymium and gold and copper and an old development of use while the table, the part of the table of mental year for the human body is very limited. And can we not develop robots with only not using those rare earth materials, but that’s also one for the future visions in the future.

Interviewer: As far as things that you’ve done so far up to this point. Is there anything that you would say is to you, you know, your most or one of the most important contributions that your work is made?

Bram Vanderborght: Yeah, I think the ability to bring different sciences together is always been very challenging but also very rewarding to see that although all the beginning struggles that it’s very impactful. And so that I think is very important that we could realize that in our university but also with with I make and with other European partners mostly in European projects, because as I like on the European funding scheme, that they really support you to work together over different disciplines. And yet we use really a lot how to combine very different disciplines. And so we also wrote a book on how, and I collected 56 academics from my university from very different fields tax lawyers and sociologists and physicists, psychologists and so on to think, yeah, imagine a society with a lot of robots and AI. What will change and what we do in doing to change can we become a friend with the robot to have taxes for robots and so on, and to think about the future, where the humanistic values on equality solidarity and freedom, how they stand ground in such a robotized world.

Interviewer: How do you think robotics has changed over the years? And I guess related to that. Do you have any ideas or strong intuitions about where you think it is going? What important developments will be in the next decade or two?

Bram Vanderborght: Yeah, it evolved a lot. Only the only the size of the robotics community. I still remember my iCRA conferences at the very start of my PhD. Yeah, we could know all the persons. I could name all the bipeds and the humanoids by heart. Now, every week there is almost a new humanoid. So the size and the investments in it is increasing a lot, which of course creates very new opportunities. But it also creates challenges on sustainability, for example. But I think we really need, although the world is falling apart in different geographical regions. I think it’s a real danger for science or human and even humanity. And I think we still need to reach out to each other and work together. I don’t think we really competitors with each other. I think the challenges of this world are big enough that we can that we need to work together to overcome them because that will be much more efficient in time of use and progress in the world in robotics had to face all those challenges. So that’s really a worry. How the geopolitical situation is now. While I think we need to create bridges over different disciplines, but also over different countries and nations had to face the challenges of our world.

Interviewer: Thank you. Well, that’s a nice transition to the last couple of questions I have for you. In terms of the world and sort of developments in the world, impacts on the world, have results of your research been applied to any real world situations? And if so, what did that look like and did it surprise you in any way?

Bram Vanderborght: Yeah, the most impactful is apart from the publications and how we impact, of course, the scientific community. I think is Axios Bionics, which is a company founded in 2019. And now they’re bionic prosthesis around the market. And we talk, of course, with users of them and they really like a lot the technology, even if they don’t know that everything the lab that was at the start of those developments. So I think it’s really nice to see persons which had before a disability that could have much better mobility due to that. And I hope that our other startups, which are much more in the starting phase, will have a similar or even bigger impact on the society. And so that we also changed it before. It was more research to have impactful papers, but I really want to push the technology further. And a lot of time now is now devoted not anymore so much to not only making publications, of course, that’s still a big part of my work, but also doing a lot of efforts how the technology can further evolve towards really implementation in society. And I did not stay in a very nice paper or a presentation at the conference. And yet it also needs a change in how you approach the research and how you plan the research in order to make it possible to create a spin-off company out of it.

Interviewer: You talked a little bit about preparing your PhD students for failure as a sort of important and generative part of the overall process. Do you have any advice to students in general, maybe younger students who are interested in getting into robotics in the first place?

Bram Vanderborght: Yeah, I think it’s very important to take risks and to fail. I also wrote several opinion pieces that you need to dare to fail. I still remember in Japan, we were letting the robot control to step over large obstacles. And we had a problem with the knee, that the velocity of the knee went over the limit and then you had to get failures. And then we brought in one of the head persons who designed the robot and he took a baseball bat and during stepping over, he kicked the knee and was like, “What?”. And he said, “Yeah, you need to push your technology to the limit and see where it fails because there you learn a lot from it.”. And so I think also companies are SpaceX, but also Boston Dynamics has really that in their DNA. It’s not that we don’t have so much money to get those amount of failures, but I think it’s important for students to dare to do things and give them freedom and guidance. It’s a balance between the two because I think in our lab it’s important that everyone is equal and that everyone comes with ideas and that we help each other out because to have such a multidisciplinary lab, we have a lot of expertise and we talk about diversity. But I think it’s also diversity in people that we hire. We have persons who had a whole career in industry and said, “Yeah, due to reasons, I want to do a PhD and maybe they’re not the experts in the latest MATLAB tools, but they come with a lot of other experience which fits in the broad spectrum of researchers we have in the lab.”. And so daring with new ideas to come and to see how they’re tuned in our lab because it’s always also a challenge, but you take us new research directions and not because I talked about success stories, but we also try things out. Yeah, that was a complete failure. And then later I thought I shouldn’t have gone in this direction because we didn’t have the expertise or the capabilities to do that. So it’s always a balance what new large research tracks we start because to develop that, yeah, it takes quite a lot of effort. And what I also then say to those researchers, I’m also very honest on it, like you will be a pioneer in this. On that topic, there is not yet a large team on it. Are you willing to do that? I will help you to build a team around it, like we have teams like the 20 persons on Selle Feeling, one that was one person like Sepu Terrean, with a colleague from the Polymer Group who started it, but they were two. And so they had to manage everything alone in their research, but then it grew and then we had expertise on simulations and on processing and so on. So the team grew, but originally it started with a single person. And that for new research checks is also a challenge, but also an opportunity to become a leading person in that field.

Interviewer: And for the last question, we like to ask everyone if there’s anything you’d like to share that we haven’t covered so far.

Bram Vanderborght: I think it is a start today to reach out to other disciplines when you’re having challenges, really try to collaborate with others. And that I like a lot in our research group that without me knowing sometimes even how they talk during lunches or events and they find matches which are without me done. And that I like a lot. That’s someone from mechanics and someone from AI, they somehow meet and they come up with very new solutions. And I think that’s very important to allow that and that it’s possible because again from there, a lot of new developments are often possible.

Interviewer: Anything more?

Bram Vanderborght: I cannot think about it now.

Interviewer: Well, this has been terrific. Thank you so much for your time.

Bram Vanderborght: Thank you and thank you for taking your time.