For today’s episode, we had a thoughtful conversation with the expert Roland Siegwart. He is a Professor for Autonomous Systems at ETH since 2006 and Co-Director of the Wyss Zürich since 2015. He is co-founder of multiple successful spin-off companies in robotics and related fields and a board member of various Swiss institutions and foundations. He is also a strong supporter of innovation and entrepreneurialism.
Elisa Muñoz: I wanted to ask you Professor, what was your first approach with robots? Do you have any childhood memories related to the topic?
Roland Siegwart: Yeah. Interestingly, a friend of mine has a memory. He told me that I only spoke about robots when I was about 10, that I will become an inventor of robots. I didn't remember this. And of course, then I really got into robotics and related fields because I have studied during this evolution of personal computers. When I started with my studies, we still had these carts for programming computers. And at the end of my studies, we had laptops, small laptops, which of course were not very powerful.
And this made me really excited about using this combination because I studied mechanical engineering, the combination between computers, sensors, mechanical engineering and robots. I think this is the most interesting field to apply these different engineering disciplines.
Elisa Muñoz: Do you remember it was the first robot you ever built and what was it for?
Roland Siegwart: So, for the first robot, I was not the main investigator, but I was involved in a robot playing ping pong.
I was there with the Institute of Mechanics where we did more and more mechatronics. So of course, mechanics and dynamics are moving parts and you want to be fast and dynamic. And then we thought, “Why not build a robot, which can actually replace one human by playing ping pong against other person?” And so it was very fast mechanics, but then we realized that we also have to, to catch the ball or actually see how the ball moves. You need a lot of modeling and then you need a vision. At this time, the computer system was about 150,000 Swiss francs or dollars only for this ping-pong playing, which today you have the same calculation power in a smartphone and you could easily do this. What we did there was these very expensive computers.
Elisa Muñoz: I did the research on your investigation and noticed that you're passionate about flying robots. So how was that transition from building a robot in order to play ping pong to actually one that it can fly?
Roland Siegwart: Yeah, actually this was quite some time later in between that also started to really use robots for education. I think it's a wonderful tool for students. So this was around 92 where we did the first problem competition with students at ETH. And I started being a professor in the 97. And so we also started to build small robots, small robots with wheels, which were the size of a sugar cube.
And then this was really fascinating, first we worked with the swarm behavior of these robots, but then we realized that smaller robots are extremely limited. They can only move on flat ground. And then we said, “Could we probably build some propellers and make them fly?” And that's where we started to do flying robots.
Elisa Muñoz: Wow. Wow. This is really interesting. Since you have so many years of experience in the field, what would you say that it's the biggest misconception of robotics? Like in, let's say in the last 10 years?
Roland Siegwart: I don't know if they're a big misconception, but I think it's extremely important that you always go also in a real field or real applications. You can do a lot of fancy robots moving simulations, but once you have to deal with real world changing lighting conditions, changing situations, surprising situations there, the challenge begins and you have to deploy. And we had all the research I did with my collaborators was always that we were then moving on a real robot and doing experiments in the real world. Sometimes it was hard. It's a, sometimes it's tiring, but once it's worked, so that at least partially works is probably much more inspiring, rewarding, because you can really see your little baby mechanical robot sensors and computers evolve in our daily environment.
Elisa Muñoz: Since you have such a background in entrepreneurship and in the field, what would you think that nowadays it's so hard for robotic startups to succeed? What do you think will be the main key in order for them to, to grow, to actually get their robots into the market?
Roland Siegwart: Yeah, it's very simple. If you start at university, you are dreaming of the perfect robot, which is probably capable of doing whatever you want. And I think it's good if researchers really have this dream and try to make one step after the other to pursue this goal, but the startups should forget about the street and should really say, what can we do today? And what can we solve, visit what we can do today? And I think there have been in the last 10 years, I think the biggest, the fastest of all the feelings in robotics was flying platforms. There are now millions per year sold for private use.
Now, why is this possible in principle? Flying is for us, seems not so easy, but this controls off. We have, we can do things quite awhile, but now once you are in the air, the navigation is much easier in the free space because you have no collision risks. No. So you're trusting the free space and you have GPS, which tells you always where you are. As soon as you're in the building, it's more difficult. And you could see that today you can buy small drones. You can fly outside, take images, but you cannot buy drones, which really fly into your home and do something in the home because they have no GPS then have to rely on vision.
Basically the technology today is ready. But it will take some time to really bring it on, on the drones. And I think it's important to, at one point, once you do research, you should go for the crazy dreams. Once you are going for the company, you should actually really step down and say, what can I do? Which can be useful and which I can sell, which visited technology today.
Elisa Muñoz: Where do you see the robotics future on the medical side? Do you see any future there or any improvement in that area?
Roland Siegwart: I think in general robots in the medical field, first of all, for intervention, it's all very well established and it's somewhat a tool. And there, of course, all the decisions taken by the medical doctors, but it helps the medical doctor be much more precise. Also our hands, we cannot hold something very strongly and precisely. These are all of us who can do a much better job. And I think this will further evolve so that we have less impact on the human body. So we do, there are a lot, these minimal invasive interventions, which are done by Vinci and all robots, then of course, robots can also help on the other side for rehabilitation so that robot people can much faster get back to normal life.
And I think in general, medicine should mainly have the view that we should. People bring back people to a normal life. It's not about giving the person at the end of their life two more months. This is probably not worse, but we have to have people back to normal life and happy life as soon as possible. And robots can help in rehabilitation, but sometimes they can also support elderly people in staying at home for longer, which is also important to have.
Elisa Muñoz: Wow, this is really interesting Professor, thank you for sharing your thoughts. Do you have any last advice for our community?
Roland Siegwart: I think in general, robotics is a wonderful feeling of adrenaline. I think young people, especially they should dream of some stuff and then they should select the most complex thing where they are afraid and they think I probably wouldn't ever do it. You should exactly go for this because then you will be excited. You will make progress. It's probably more tough than you think, but I think you should always go for the big challenges. And robotics has a lot of wonderful challenges, a lot of different fields where you can learn and grow. And hopefully I make a lot of progress for society in the context of climate change and all these issues.
We will run at the limits of feeding the whole world's population. And I think technology robots will help us hopefully to be much more efficient on the field, but also less have a negative impact in the field. For example, not spraying the fields with chemistry and pesticide, but just bringing the goal very precisely with robots and interfering rates needed.
Elisa Muñoz: Once again, thank you professor. It was, it was our pleasure.
Roland Siegwart: Thanks so much.
This interview was brought to you by ControlHub, the most intuitive purchasing software to request, approve and track all your business purchases.