Germany is among the world leaders when it comes to the use of industrial robots. Over 230,0000 are used in the national economy. Germany thus leads the European comparison. But in society, many people have other ideas in mind when they hear the word "robot". Our impressions are shaped by movies and series. C3PO and R2D2 from Star Wars are perhaps the best-known examples. But humanitarian robots are far from being dreams of the future from science fiction. We talk about the use of robotics, the limits and the societal changes in a new episode of our science podcast.

Today's guest:

Dominykas Strazdas is a research assistant at the Faculty of Electrical Engineering and Information Technology, in the department of Mobile Dialogue Systems. Under the supervision of Jun.-Prof. Ingo Siegert and Prof. Al-Hamadi, he works on human-machine interaction, among other things. His team has a very strong international focus. PhD students from Egypt or India, students also work on their project work in the Robotics Lab of the Department of Neuro-Information Technology.

*the audio file is only available in German

The Podcast to read

Intro voiceover: Know when you want. The podcast about research at the University of Magdeburg.

Peer Niehof: What 50 years ago was still a dream of the future has long since become part of our everyday lives. Today, we encounter robots on a regular basis. Often we don't even really notice them because they seem so familiar. Or would you perceive the chat robot when shopping online or the cleaning robot in the bathroom as a classic robot? Probably not. Instead, we immediately have a different image in mind when we hear the term.

Often also shaped by movies and series. Millions of robots are in use around the world. They roll, walk, talk, work, labor and help people. They take on dangerous, difficult, but also monotonous tasks for us. The technology is developing rapidly. But is there no limit? And doesn't this development also bring threats to our society? That's what we're discussing today on "Know when you want."

My name is Peer Niehof and I work at the Otto von Guericke University Magdeburg in the field of media, communication and marketing. Today I would like to welcome Domas Strazdas to our podcast. He is a research assistant at our university and deals with the topic of robotics. Hello Domas, nice to have you here.

Domas Strazdas: Yes, hello!

Peer Niehof: Domas, when did you use a robot today?

Domas Strazdas: Well, today I must confess not yet. Which means, I must confess. So I've been thinking about whether an electric toothbrush would pass for a robot. But I would say not yet, because it doesn't somehow autonomously explore the world. But a little vacuuming robot, I'm going to use that tonight. So in the kitchen, in the evening when we're done, sucking away all the crumbs after dinner, it's super annoying and it does it pretty well. So today we have the first.

Peer Niehof: I was already listening to music in the shower this morning via a smarthome assistant. Does that already count as a robot?

Domas Strazdas: Um, it doesn't have a physical body. I'd like to separate that somewhere. Can we also discuss a little bit later what constitutes a robot? There are also the bots and the robots. And the robots have a solid body and the bots are anchored somewhere digitally. They also have something in common, i.e. how they act or how they find their way in the world.

So the bot in the digital world, the robot with us in the normal world. And both perform tasks that are mostly a nuisance for us. That means they also have a lot in common. But for a real robot, I would say, there has to be something to touch somewhere.

Peer Niehof: Okay, then that didn't count, I guess. When we talk about robots, each person has a very different image in mind, often shaped by movies and series. Some immediately think of R2D2 from Star War's or C3PO. Others may have gripper arms in mind that assemble our cars in workshops or large car factories. So what types of robots are there? Is it possible to say in general?

Domas Strazdas: So you can classify robots in many categories, depending on the purpose, meaning field. I'll say there are industrial robots, service robots, medical robots. But you can also classify them by appearance. There are humanoid robots, zoomorphic robots that look like a dog and walk on four legs. But you can also classify them by stationary or mobile. And most robots have a couple of those categories to serve a particular purpose. And a couple of the categories are more common. Together, for example, service robotics, mobile robotics, and humanoid, I'm sure you have an idea right away. And then: Industrial robotics, stationary, and then precision tasks, that kind of thing.

Peer Niehof: And which of them do we encounter the most in everyday life?

Domas Strazdas: It is difficult to say directly. We get a lot of products from the robots. For example: most cars are made by robots. In that sense, we encounter the products from the robots. To really see the robots in everyday life, if you don't meet any service robots, is not yet the case, I would say.

Peer Niehof: Okay. But that sounds like a lot of types that we can theoretically encounter or that are present in society. But in general terms, what kind of technology do all these robots actually consist of? Are there any common denominators?

Domas Strazdas: You can think of it a little bit as a robot being a PC on wheels or feet. Ultimately, it's a small, smart machine that either moves around in the world or performs a task in it. That means there's a mechatronic system, so actuators that move this robot along or make it move. Multiple interfaces most of the time. Then we have a controller for those kinematics, but also the general controller for the programmed sequence.

Then we have sensors to perceive the world to perform this task. These are mostly cameras or rangefinders - lidar systems. Basically, a self-driving car could also be considered a robot. Then we have, as with a laptop, a rechargeable battery to power it, and you have a robot that can do something.

Peer Niehof: Let's stay with the humanoid robots, because this is perhaps where many of our listeners have the most images in their heads. Do they already come close to their role models from the well-known films and series, or are there limits?

Domas Strazdas: So humanoid robots, if they are not used in the reception area, are occasionally also used in the catering area. They come up to you and bring you something or ask if you want anything else. And I have encountered such a robot before, albeit in a somewhat reduced form. In a hotel in Munich, I was able to order this robot by phone call.

That is, then a robot voice came to meet me, I pressed a few buttons and then he drove up to the door. I got another call on the phone: "Hello, I'm here!". Because he couldn't ring the doorbell yet. And then I opened the door, he stood in front of the door and greeted me and I could then choose on the display what I would like.

And then I ordered a bar of chocolate and a beer and then pulled out of a drawer. I was able to take that out and at the end I was allowed to rate the whole interaction. He got five stars from me, of course, and then I took another quick look around. So through the door and there I saw in the hallway, a person who followed the robot with a controller, that is, the whole interaction.

And I was just interested in whether all of that really happened for real or whether it was perhaps, let's say, "faked". And when I saw the person with the controller, I waved him over and said, "Yes, let's talk a bit about robots," and he said, "Yes, of course, that's our prototype right now, and then I installed a new software so that he can ride on elevators on his own. And since I'm not sure if it's all working properly, I have this controller just in case, so I can activate that. Not that the robot is going to roll down a flight of stairs." And then I was assured that the interaction was then ultimately real. And because it was only a prototype, I even got my beer and chocolate bar for free.

Peer Niehof: Very good. A win-win situation for everyone. Can you describe what that looked like? Did it come close to that goal of being human?

Domas Strazdas: Well, it was humanoid in that it had a head, but it didn't have hands, so it didn't need them. Instead, it had individual drawers that it could move
out, depending on what was inside. Basically, it was a refrigerator on wheels, so it had a head and a touch display.

Peer Niehof: Are these also the limits that these humanoid robots have right now? The appearance or where else are the limits?

Domas Strazdas: There is a special kind of humanoid robots that look quite human-like. These are called androids in common parlance. One also encounters them more often in science fiction. Then when a human sort of turns out to be artificial, completely artificial, that's an android. The hybrid form between humans and robots. One would call it a cyborg, but that's another topic.

An android has nothing biological about it, except its appearance, and that can also create discomfort by looking too human-like. And you get that and then you feel a certain distance because you realize: Oh, that's not a real thought now, not a living being. Pretends, but as if it were a living being. And in order to reduce this effect with the service robots, they don't dare take this step towards this total android appearance.

And the robots are made to look robotic. That the person who approaches it knows from the outset: Oh, this is a robot that has perceived me, wants to interact with me, I will also interact with it. But like with a robot. So some people, as soon as they know that they are talking to a computer, then adjust to certain things, as they also know from television, how you talk to a robot.

So nowadays I would say this communication is also influenced a lot by interaction with smartphone. There are so many voice assistants and ultimately the same channel, the same process would be used with a robot.

Peer Niehof: For many listeners, this may all still sound a bit like science fiction and everything may also be a bit far away. But you've already given an example. Are there other areas of application where humanoid robots are already being used?

Domas Strazdas: Except in service and gastronomy?

Peer Niehof: Yes, exactly, or perhaps also in prototypes. So what is possible later on, what is perhaps being tested and tried out.

Domas Strazdas: Such service robots are also used in the medical sector to partially animate people to perform certain movements. So it's the same for retirement homes. There are small robots that dance and the seniors dance along with them, and most of the time the seniors actually like that the small robot dances. And the relatives have their reservations. But when they see that the whole interaction is going well and that the people are happy, then they have fewer concerns.

Of course, there are also many prototypes, especially in the area of research, the next generation, which is being bred, so to speak. There are also many humanoid robots that are being tested to see what their possibilities and limitations are.

Peer Niehof: A second important area. Then let's move away from humanoid robots is, of course, for business. Right now it's industrial robots. According to statistics, the most of them are used in Germany throughout Europe. Even in a global comparison, Germany comes directly after the Asian countries. Why is Germany so far ahead in the use of robotics?

Domas Strazdas: This can also be explained, I think, quite simply with industry. Robotics was the direct response to automation and the next step in industrialization. And Germany, as the world's leading exporter, also helped to breed these technologies. A good example would be Kuka.

Peer Niehof: Mhm, can you explain what that is?

Domas Strazdas: Oh, I would have to look up the names Konrad and something Augsburg. So A stands for Augsburg and now I'd have to look up what the namesakes were.

Peer Niehof: But what they do. What is the name Kuka?

Domas Strazdas: Kuka is another success story from Germany. They started out with industrial welding systems, well, they didn't start out, but they became really big with it. And then they continued to automate until they finally realized that a robot with a welding arm can do pretty much any task, and they were really far ahead of their time, and then they took huge steps in this direction with the German automotive industry, breaking records over and over again.

They also created the first robot, a Kuka Titan, that can lift a ton. It was previously unimaginable that a robot could lift something so heavy. And nowadays, an industrial robot can theoretically lift and turn an entire car body as needed. That would not have been possible with conventional technology in the past.

Peer Niehof: So that sounds logical, of course, when you say: okay, it's become big in Germany because of the automotive industry. But are there other areas in which these robots are used?

Domas Strazdas: Ultimately, these robots are part of industrialization and part of automation. So you can palletize, you can stack, you can do quite a lot in logistics with it, which makes the dull and repetitive work much easier. I don't want to stand in the fish factory all day and put the fish in a can. A robot is just the right candidate for that.

Peer Niehof: This may explain why robotics is so successful in Germany as an industrial nation. But we had heard that Germany is actually only in the top positions, but still behind Asian countries. Why are the robots or why is robotics so, yes, so well-known in the Asian countries? And above all, is the technology being driven forward in this way?

Domas Strazdas: They have a somewhat different cultural connection to it, i.e. they are more present in the media, they have a somewhat higher acceptance, I would say. Ultimately, when it comes to the hard numbers in the industry, Germany doesn't have to hide at all. So it's not a factor of two or three, but maybe 10% behind.

Peer Niehof: But that's true. If you always somehow notice the latest developments in robotics, in the media, in the news, then it very often comes from Asia. I always have that in mind myself when I think about robots or when I have pictures in my mind's eye, that this development somehow always comes quite often from Asia. That was just very interesting.

There are robotic servers, robotic packers, robotic garbage collectors, and even forklift trucks that can now operate without humans behind the wheel. This begs the question: isn't this a real job killer?

Domas Strazdas: You have to look at it a little bit more differentiated. Job killer I say technology or technology change - is it's a job killer when you look at the old jobs? But it's a change. With new technology, new jobs are being made and the old jobs are being changed. So if the weaver of the past made one piece of fabric, one carpet, but then today he would make hundreds of them with a weaving robot, with a weaving machine, without having to do hard physical labor now.

These robots take over the really dull, repetitive, non-ergonomic tasks and that simplifies the human's job. This means that the human being then stops doing this physically strenuous work and then starts to monitor this robot, to provide the ideas, to equip the robot, to maintain it. There are many areas, so this process does not run automatically just because a robot is used.

So the person who used to operate this loom or still did everything by hand can start to take care of more advanced tasks.

Peer Niehof: But even so, some industries are already a bit afraid that they will die out or, let's say, be displaced by robotics. How do you see that?

Domas Strazdas: That will happen in part, that certain tasks will be taken over. But that doesn't mean that you can't specialize further. So craftsmanship will always be in demand. There are people who simply appreciate the aesthetics of human work and want to do that, and a master carpenter will still be able to deliver his masterpiece, and a machine from the factory will never come close to that.

Peer Niehof: In preparation for the podcast, I naturally did a bit of research and read a few interesting papers, and I got stuck on one. The paper discussed whether robots can also be used as managers. After all, technology is so far advanced that it can take over personal and social interaction. It was a pretty detailed paper and I think there's a lot of science on this already. So will we soon have to prepare for a robot boss or a robot supervisor?

Domas Strazdas: So I think you're just referring to a study from MIT and there, human and human-robot hybrid teams have worked on certain tasks, mostly to put Lego together. And what they found is that an autonomous robot that also gives tips on how best to move forward as a team, that that was found to be the most pleasant. So manager is maybe hard to say, so he didn't stand and say: Man, do this, man, do this and all the people like that: Yay, we are so happy about this.

That was not the case. So ultimately, you can think of it as an algorithm that is optimized to do a task most efficiently will do it better than humans. That is, if this robot is concerned with plugging these parts together as efficiently as possible, it can keep pretty good track of it: Okay, human A is just putting these pieces together, human B is just putting these pieces together.

So it would be best if I now take over this task area C and then bring the other components to person A that he needs. So now another person comes along and asks: What should I do? He already has this sequence in his head and says: Okay, then you do a second part B and then we're done with all the tasks, roughly at the same time, optimized like that, you can do that.

Of course, a real manager has completely different areas of responsibility. I would like to see this robot in crisis management. I don't think it can do anything at all at that moment. Even promotion, social interaction, is still a long way off. So robots won't be replacing managers for a while yet.

Peer Niehof: What then actually applies better as a summary: humans will be replaced or our understanding of work will simply have to change fundamentally in the future?

Domas Strazdas: With the change in technology, which has always been there, people have always had concerns. Ultimately, through the various interactions of industrialization, that is, with the steam engine, with electrification, with assembly line work, that is, line work, we have become more and more efficient. And through that efficiency, it must be said, our available leisure time has also increased. So the average person has to work much less than 100 years ago or 200 years ago.

I don't have to go out into the field now and harvest my potatoes myself. The farmer does that with a super large tractor, partly automatically GPS controlled. He types in a few coordinates on his cell phone and the thing drives off. And that allows us to cultivate large fields, which would not be possible otherwise. Or if I now imagine fruit harvesting, that was partly like that, you drove and had a device that drove through the area, then there were ten people lying there or partly still lying there, and while the thing drove by, they picked some piece of fruit or vegetable from the ground.

And that in wind and weather. Now that's a job I wouldn't like to do. And when I consider that a robot can drive through the area and also press weeds into the ground somehow with such small sticks, because it has recognized that automatically. And then it can also pick the weeds. That's much more pleasant and people can do more pleasant things, I'd say.

Peer Niehof: So we can at least state that not all people welcome this development and are at least a little afraid of possibly losing their job. But there is not only the fear of losing one's job, there is also a so-called robo-phobia, i.e. an aversion to robots. Do such fears also play a role in technical development, for example how robots can have a more positive effect on the human psyche?

Domas Strazdas: Yes, definitely. So we noted earlier that a certain discomfort is emitted by the robots when they appear too strong and Android, too human-like. This effect is then called "Uncanny Valley", meaning creepy moat or discomfort valley and this valley suggests on a graph that on one axis that represents human-like, on the other axis the positive feel.

Then, the more human-like an object becomes, the better that is felt. Until there is a critical point, where suddenly this curve sinks rapidly into the ground and partially the positive, the positive axis passes through, so it is really a negative feeling, until that increases again to the conclusion to the perfectly human-like object then again. So as an example of this effect a stuffed animal has already human-like character, a teddy bear has the poor proportions differently than a real bear.

And we feel positive about it, because we are happy about it, we say: oh, it looks cute. But when these robots are built in these types and at some point are made more human-like, for example with silicone masks, that the whole robot body is disguised, then it almost looks like a human. But you suddenly realize: Oh, that's not a human at all.

And then an uneasiness emerges. Perhaps one could also explain this. A lifeless human body, i.e. a corpse, would cause exactly the same effect. But if these objects also move, then this effect is amplified. So a corpse in a room causes discomfort. But if this corpse were to move, i.e. a zombie, then it is the most uncomfortable thing a human being can feel.

And exactly the same thing happens when we have this android robot, which is clearly a robot and not a human, so it's kind of a corpse, starts to move, then you get really scared. So that's why this effect and how to get around this effect, that's a big question in research. And we currently have humanoid robots that we also want to investigate in this area, how they affect humans.

This is also partly a psychological question and currently we want to investigate the voice, for example: how does a voice have to sound so that it doesn't trigger this effect? Because synthetic voices can also have this effect when you realize: oh, that's not actually a real person I'm talking to, and the expectations must also match. When I talk to a robot, I actually expect it not to be a human being at first and then to give me a robot-like answer.

And if I suddenly hear a real human voice, how do I feel about that? And it also plays a role whether certain gender characteristics can be assigned to the robot. Because in the voice, we also have a spectrum between male and female. And where does the robot voice have to be located? And should it actually sound robotic? Should it sound male or female? We still want to find out all that.

Peer Niehof: Because you just mentioned that, also with the teddy bears. Is there a different perception that children have towards robots than adults? Because a teddy bear is perhaps for children, it stands for something cuddly. They like to hold it in their arms, they like to hug it. But would it be similar with robots or do children react differently?

Domas Strazdas: So they would also find their way around in this Uncanny Valley. But they would perhaps show it even more clearly. So I took my son into our robotics study. He is three years old and he is actually afraid of the one humanoid robots. He said, "Dad, please turn it off. And at that point, I have to say, the robot looked even a bit scarier because it had funny eyes, I would say, and that is a very important criterion for a robot that looks human-like.

What effect it has on you. And then the robot started to look around, of course it also noticed the human being and that was too much for my little one at that moment. And then I switched off the robot at the last moment, when it collapsed almost lifelessly. The eyes went out, the arms down, was the biggest moment of shock. And then he didn't even want to go near the robot.

Peer Niehof: But then let's get more into your research or your research here in Magdeburg. I've been to your studio before, to your lab, not to the studio. And when you're there for the first time, it feels a bit like a little playground for adults. Can you describe to our listeners what it's like in your lab?

Domas Strazdas: Yes, definitely. So we have a stationary robot that is attached to a table and can handle cubes. So it can stack them, which also seems playful, but offers a lot of room for interaction, so you can give commands to this robot. You can also say: give me this cube, build a pyramid, spell a word, because there are still letters on the cubes.

This is an interesting field, because you can also use camera technology, find these letters again and start the interaction in many different ways. In addition, we also have two stationary robots, one with a gripper arm that can be used in many different ways, that can also pick up various objects, even from the floor. Besides our stationary robots, we also have two mobile robots that can orient themselves in our laboratory.

These robots use a so-called slam algorithm, i.e. simultaneous localization and mapping. That means they locate themselves and map the space simultaneously. And that has a lot of challenges. When you start moving into a space, you have to find certain features to locate yourself in it. And you have to recognize these features when you arrive at the same place.

And then the whole thing has to be combined with the rest that is interactively present in this space, i.e. interesting objects, which then also have to be recognized again and correctly located there. Then, of course, the people must also be recognized and that the people can be approached, perceived. And if the human gives a certain task: Man, robot, go there, do something - then he must also plan the path correctly.

So there are other challenges. Both stationary and mobile robots interact with people by perceiving them via cameras. The cameras usually also work in three dimensions, which means that you also have deep data. To interact with the robots, you need sensors, and in addition to the cameras and microphones installed in the robots, we also have other cameras and microphones stationed in the lab, and in some cases microphone arrays, i.e. microphones with many smaller microphones next to them.

And they allow us to use delays in time after a sound happens to determine where the source of the sound is. That is, if we see a scene with many people and one of them is talking, we can then also determine which of the people is currently saying what - not only visually, but also acoustically. And then they start to extract all the features that are interesting for us and the interaction, i.e. gestures, facial expressions, speech, attention.

We want to find out whether the human actually wants to interact with the robots and whether his attention is on the robot. And we can then determine whether he looks bored, whether he is afraid, whether he is happy. The whole thing can already be determined via the cameras. And then, of course, we want to do something useful with all the data we have collected.

That means we want to start a corresponding reaction of the robot and then see how people respond to it.

Peer Niehof: All this, and especially the lab, is only possible because you work at the Faculty of Electrical Engineering and Information Technology. How did you get there?

Domas Strazdas: Yes, I started studying mechatronics. That's already going in the right direction. And after graduation, I started a doctorate in neuro-information technology. This is a department headed by Professor Al Hamadi. And during my doctorate, I spent almost four years in the robotics department of a large automobile manufacturer. There I gained the industrial experience that is needed.

And now I am anchored in the department of mobile dialog systems, headed by Professor Ingo Siegert, and the two groups of mobile dialog systems and neuro-information technology are doing a project that we call "Rosa" - Robot System Assistant. And there, the competencies of the two departments can come into their own very well. This is my project, so to speak, my affair of the heart, which I will ultimately also write about in my doctorate.

And all of the individual areas listed earlier are represented there. That is, we have attention recognition, we have speech processing, we have gestures, we have facial expressions, we have mobile robots, we have stationary robots, and we also have a large working group that takes care of these individual areas. That is, we are now two professors and seven doctoral students, each of whom has chosen one of this area for their own doctorate, their own in-depth view.

And all of the individual areas listed earlier are represented there. That is, we have attention recognition, we have speech processing, we have gestures, we have facial expressions, we have mobile robots, we have stationary robots, and we also have a large working group that takes care of these individual areas. That is, we are now two professors and seven doctoral students, each of whom has chosen one of this area for their own doctorate, their own in-depth research.

Peer Niehof: When you leave our podcast recording studio right now and head over to your lab, what will you be working on today?

Domas Strazdas: Currently, we have started to better connect the mobile robots with the stationary robots and for this we have a so-called "April Tag". This is a somewhat different QR code printed out. And these tags can then be seen very well by the cameras, even from a distance, and then a 3D system is immediately projected onto these tags. That means you know exactly the orientation of these things.

These tags are used in part by NASA, so that's top technology for it now, and we thought that would simplify our configuration. If these robots would come together in a common world with the help of these objects.

Peer Niehof: So we learned today in the podcast that we actually encounter robots everywhere and the leaps in technology in recent years and decades have been truly remarkable. But let's take a look into the future. Very briefly and succinctly Domas: how should we imagine robots or robotics in ten years?

Domas Strazdas: So there will definitely be more robots, smaller robots that will take over the tedious tasks at home. So there will be robots that mow the lawn, there will be a robot that vacuums away the dust. There will also be robots that clean the windows. These robots will gain popularity, especially in homes. There will then also be the individual robot probably in the restaurant sector.

So at McDonald's, you can imagine that somewhere there will be something to see, so as examples.

Peer Niehof: Okay, maybe when we meet again in ten years, we'll say: we already used robots this morning. (laughs)

So I'm very curious to see whether all this will come true or perhaps it will be quite different. We will keep an eye on it and also keep an eye on the research here in Magdeburg. Domas, thank you very much for the interview and thank you for being here.

Domas Strazdas: Yes, thank you very much. It was fun for me, too.

Peer Niehof: If you're interested in learning more about research at the University of Magdeburg, I invite you to listen to the latest episodes of our podcast - if you haven't heard them yet. They deal, for example, with the question of how the EU will master crises in the future or whether hydrogen is the future of our mobility. We are always happy to receive your feedback, be it on this episode or on other editions, by email or on our social media channels.

We will be back for you in two months with a new episode. Until then, ciao, bye and stay healthy.

Intro voiceover: Know when you want. The podcast about research at the University of Magdeburg.