Interview Transcript
Q1: So we like to begin by asking you a few background questions. If you would share your full name, date of birth, place of birth, anything you want to share about your family or early school.
Hyouk-Ryeol Choi: So my name is your choice. I born in 1962. Well what else? What else? I have to do?
Q1.1: Anything that you feel like sharing that might be relevant about, you know, or experiences in relation to your career?
Hyouk-Ryeol Choi: Yeah. Okay. So I’m truly educated inside of Korea. So I got my bachelor’s degree in Seoul National University and the master’s degree in the Korea Advanced Institute of Science and Technology. That is called KAIST in brief. And then I got my PhD degree in the Pohang University of Science and Technology. In between the master’s and PhD study, I went to a company that is called LG Electronics. I spent four years as a member of—well, it’s a little bit different—but I’m working on the development of a laser beam printer. That was my old career.
Q2: Are you still with LG now?
Hyouk-Ryeol Choi: No, I quit.
Q3: What’s your position currently?
Hyouk-Ryeol Choi: I’m a professor in Union University. It’s a private university in Korea, but it is supported by Samsung Electronics. It’s a private university.
Q4: Did you have any mentors during your school career that were significant that you would want to mention or talk about?
Hyouk-Ryeol Choi: Okay, so well, I just had a quick meeting with my previous advisor at this conference. I didn’t know that, but he came here. I was very, very surprised. Well, he is very much older and retired, but he still has a strong interest in robotics and robotic research.
He visited this conference. I met him, and then I talked about the current issues of robotics. But he’s a real mentor of mine. Well, actually, up to my master’s program, I didn’t have any experience in robotics, but I started my career in robotics from the POSTECH PhD program.
Q5: You might have already mentioned his name, but if you didn’t, can you share who your mentor was?
Hyouk-Ryeol Choi: Well, it’s no problem. His name is Young-il Young. He’s a very famous professor in Paris.
Q6: Thank you. So how was it that you first got interested in robotics?
Hyouk-Ryeol Choi: Well, frankly speaking, that was at the starting of my PhD program at POSTECH. My advisor at that time handed over a thin book. The name of the book is Hand. Yeah. The book is about the biomechanics of the human hand. And then he told me, “Oh yeah, Mr. Choi, let us start studying the robot hand, which is very similar to the human hand.” That was a very long time ago, but I didn’t know that there was this big issue in the current age. Many, many people are working on the robot hand recently.
Q7: Were there any—do you remember the first robot that you ever encountered, whether it was in media or in a book or anything like that?
Hyouk-Ryeol Choi: Well, that was a direct-drive arm developed by CMU. That was called the DLR? Yeah. It has six motors. The motor is attached to the joint directly. I cannot remember the name of the professor who developed the robot arm, but there was a kind of big issue at that time.
Q8: Had you seen robots represented in movies or TV or anything like that before that point?
Hyouk-Ryeol Choi: Well, that was not a real one, but just a kind of industrial robot arm. That’s all I had encountered.
Q9: So the first robotics project that you worked on was during your PhD research on this hand. Was there anything more that you wanted to share about that specific project?
Hyouk-Ryeol Choi: Well, yeah. Robot hand. The major issue was how to actuate a multi-degree-of-freedom robot hand, and especially fingers. Most people were working on the tendon-driven system. Especially Japanese research had very good tendon-driven systems. And then I met a very good professor. His name is—I cannot remember his name. He was very good at the tendon-driven system. That was a major issue, but I don’t believe that is the only way to drive the robot hand. So I tried to find a different way to do it. Yeah, it was not easy.
Q10: What were the next projects that you worked on for robotics after that one?
Hyouk-Ryeol Choi: Well, during my PhD study, I was just working on robot hands, several versions of robot hands. One of them was direct drive, a small robot hand, using a small motor and direct-drive technology. And also, I eventually graduated from the tendon-driven system. Unfortunately, yeah, tendon is the only way to go? But I was not satisfied with that achievement because, you know, when the hand is very small, I have to have a very big driving system on the backside of the robot hand. That was the reality in that case.
Q11: Did you say you were with LG between your PhD and your postdoctoral work? Is that right?
Hyouk-Ryeol Choi: No, between my master’s and PhD. In between the two, before POSTECH, I had to serve my military obligation in Korea. So instead of doing real military service, I had to do some private company work. I worked at LG.
Q12: So after your doctoral work, did you continue to work on robotics immediately after that?
Hyouk-Ryeol Choi: Well, the thing is, during my PhD, I was very fortunate to get a scholarship from the Japanese Monbusho. And then I had a chance to go to Japan during my PhD studies. That was almost at the end of my PhD. I moved to Kyoto University. At that time, my professor was Tsuneo Yoshikawa. He was a really famous professor at that time. I went to the Yoshikawa lab and spent 1.5 years as a student. At that time, I continued my finger, hand, and grasping-related research there.
Q13: Was that postdoctoral work?
Hyouk-Ryeol Choi: Well, that was still during my PhD. It’s harder to define. I went over to Japan during my PhD and postdoctoral period, but it overlapped. It was at the end of my PhD and the start of my postdoctoral career. Anyway, during that time I spent the money from the Japanese program and stayed there.
Q14: Okay. What are some other major robotic projects that you’ve worked on over the course of your career?
Hyouk-Ryeol Choi: Well, during my study, I focused on the robot hand, and additionally I did some work on the application of robots to harder situations. For example, the POSCO Pohang Steel and Iron Works. At that time I helped with some projects. I was not the major person; I just helped with some things, for example handling very heavy blocks or heavy objects. In that case, I was kind of secondary support. That was my role at that time.
Q15: Any other major projects or work that you would like to talk about right now?
Hyouk-Ryeol Choi: Well, my good projects started from my career as a professor at the university. I did many, many projects over the last 30 years. It is harder to describe all of them, but mainly they are classified into three subjects. One of them is professional service robots. The other one is soft robotics. And the third one is interaction control. Those are the major themes of my work in Korea right now. In professional service robots, I did most of my career on developing in-pipe service robots. Maybe the robot that is utilized for moving inside a pipeline—very small, narrow, and long ones. In that case, the application is very wide. For example, gas pipelines, sewer lines, or heavy industries. I did many, many projects on that kind of service pipeline robot subject. And then wall-climbing robots and even climbing robots. And even soft drones. I developed many, many robots. Also, the paper on in-pipe robots has a very high citation. It is over 600. I think that is one of my most recognized studies among robotic researchers.
Q16: And also soft robots. This is one of your important contributions, I think.
Hyouk-Ryeol Choi: A long time ago, many people did not have much interest in material-based research for robots. For example, material-based sensors, material-based actuators, and so on. But when I looked at the other side of the study, many people were working on intelligent or smart materials. I joined that group and found that there was a very good possibility to apply this technology to robots. So by using that, I developed new actuators and developed new robots actuated by smart materials. That was actually the basis of soft robotics. Recently, many robotics researchers have become interested in soft robots similar to that approach. But unfortunately, I closed that study recently because I found that it is a very good subject for writing papers, but it is very hard to bring it into reality for real applications. It is very hard to make money with that technology. So I shut down my lab on that. I spent a lot of money building that lab, but unfortunately I had to shut down the lab.
Q17: And the third issue is interaction control.
Hyouk-Ryeol Choi: Recently, many people are talking about physical intelligence. In physical intelligence, I believe that the most important thing is sensing—the physical interaction is important. So sensing force, sensing torque, sensing touch, and sensing proximity are very important. So I decided to develop those sensors. I developed force sensors, force-torque sensors, proximity sensors, and other kinds of interaction sensors. Then I brought them into a company. I built up a startup. The company name is Aiden Robotics. It is now exhibiting at this conference. The company is developing force-torque sensors. The specialty of my sensor is that it is very easy to use. The second point is that it is robust in contact. You know, physical interaction control sensors have many different environments, but recently many sensors are very fragile in contact. There is a problem. So I made a strong, robust sensor. And another important point is that it is cheaper. I put together many conveniences for the user. It is very easy to use. You just plug it in and you can get force data. So I built up the startup called Aiden, and we are working on providing any kind of contact and physical-contact sensing modality to users.
Q18: So I did two major areas, but I mostly focus on the physical interaction control issue.
Hyouk-Ryeol Choi: There is a very good opportunity for making money recently. Many people are trying to buy my product, our product. We are developing sensors, and many people think that if you develop sensors, you are very good at controlling force and torque. They naturally believe that. So I tried to build applications using my sensors. For example, finishing solutions, grinding, polishing, sanding, and grasping. Manipulation is one of the biggest issues recently. So I developed the sensor and developed application solutions. We are extending the usefulness of the sensor to many areas, not just robotics. We found that this sensor technology is useful in other areas, such as medical applications, logistics, manufacturing, and many other fields. For example, in the medical field, implant technology—such as replacing teeth. In that case, the company needed to simulate the stress situation of the patient. They needed to sense the force for the pre-development and recovery of teeth. Some tuning was necessary. So some companies contacted us, and we built a force-sensing setup for implanted teeth. We built it using five or six sensors, added some software, built the UI and UX, and they liked it very much. We are planning to exhibit next week for implant societies. It is totally different from robotics, but we found that the technology is useful. I think we use many learning technologies recently, but they mostly learn motion data. In addition to that, if force and contact information are added to learning, I think it will be very useful for producing real human-like motion and human-like touching and manipulation technology.
Q19: So was the majority of the work that you just described something that you are pursuing through your startup or your business?
Hyouk-Ryeol Choi: Sure, yeah.
Q20: What is the name of that business?
Hyouk-Ryeol Choi: Aiden Robotics. Hyouk-Ryeol Choi: The meaning of that name is AI-driven component and robotic system.
Q21: Is the work, I assume, also connected to research that you’re pursuing at your current university?
Hyouk-Ryeol Choi: Yeah, yeah. I am still a professor at the university, but many of my students moved to the startup and they are working together. I still have graduate students in my lab, so there are plenty of applications in this context. A lot of applied research is being applied directly.
Q22: Yes, very nice flow.
Hyouk-Ryeol Choi: Yeah.
Q23: So you’ve shared a little bit about the projects you’ve worked on. You may feel like you’ve covered this, so we can move on. But is there anything in terms of research questions specifically that you want to share about questions or problems that have motivated the research that you’ve been undertaking for the past many years?
Hyouk-Ryeol Choi: Well, my background is in mechanical engineering. Therefore, I mainly focus on how to actuate systems and how to make them practical and bring them into practical applications. To do that, we need high performance, small size, low cost, and other characteristics. So I mainly focused on the actuation issue. That was the driving force for looking at smart materials, sensors, and actuators.
Q24: What are some of the big challenges that you’ve experienced while you’ve been doing your research? And that can include things that are technical challenges, institutional challenges, or even social challenges. So I leave it to you to decide if there is anything relevant—any kind of events or circumstances that have affected your research in this way.
Hyouk-Ryeol Choi: Well, as a professor, I thought about building a startup several times. I mean at the age of 40, at the age of 50, and well, now I am close to the age of 60. In my 40s and 50s, I thought, “Is it possible to build a startup?” Because I did many, many researches, but what happened after doing research and writing papers? Nothing changed around this. Even for my wife, she told me, “You are doing many things. You are writing papers, but does it make money? What happens? Many people read your papers, and then that’s all.” So I thought about starting a company, but I didn’t do that because there were many relevant problems, especially money, people, and time. But when I was close to 61, one of my graduate students who got his PhD in my lab visited my office and told me: “Professor, you have so many achievements—sensors, actuators, systems, and many other things. I think it is time to bring them to the real world. Let’s get together and do some business.” He told me that at the time. I hesitated a little bit, but I told him, “Okay, let’s start.” After that, he came to me, and two other students joined us. Eventually, we built up the company with four members. I am very happy about that because they each have their own specialty in business. We have a very well-organized group for managing the company. After that, many of my students in my lab thought, “Oh yeah, this business may be promising.” So from that time on, they tried to join my lab and my company. Now we have almost 20 PhD students in my company. So I think it is a very nice chance to manage the company because we have good human resources for driving this company. There were not many big difficulties, but I had several chances to think about startups and business, and finally I started a real one and began managing it.
Q25: Have there been any tools that have shown up over your career that have had a big impact on creating new possibilities for the work you’ve been doing? And this could be hardware things, whether it’s something like the Kinect or a new kind of actuator or something that appeared in the market. Or it could be things like LMS and particular kinds of software. Is there anything significant in terms of tools that have shown up?
Hyouk-Ryeol Choi: Well, in robotics, I think the most effective way for showing my work to the public eye is through real hardware. I think that might be the most important thing. Now I’m doing this interview, but I’m not good at this kind of situation. So we try to build robots and then show them to the public as much as possible. By way of some SNS or other methods, we show our research and our products to people.
Q26: So let’s talk a little bit about funding. Who has been responsible for funding your research over the years?
Hyouk-Ryeol Choi: Initially, I received small funding from the Korean Government Science Foundation, which is driven by the Korean government. Then after that, my reputation increased and I had a chance to get funding from private companies. So it has been half and half.
Q27: How did you develop those relationships with the private companies?
Hyouk-Ryeol Choi: The thing is, I have to show them what I’m doing and what achievements I have made in my lab and my company. So I have to go to meetings, conferences, and workshops, and show that this is my work and this is important. I have to give many speeches and talks. That is the only way to get interest from companies or other related people. That is important.
Q28: Would you say that the funding that you’ve received, whether where it came from or what type of funding, has had an influence on the research that you’ve done or been able to do because of it? Or failed to be? Have you not received funding for things that you would like to have pursued but haven’t had a chance to?
Hyouk-Ryeol Choi: Oh, that was a really hard memory. I think most countries have a similar situation. You know the meaning of Engineering Research Center, abbreviated as ERC in Korea. There is a program driven by the government for almost 15 or 20 professors altogether, doing the same research together. It is a really big program. I applied for that project. I spent almost six months preparing that proposal, but I was the last one. I failed. The most severe thing was that I was defeated by a younger professor at another university. I thought that I could be the winner, but the result was not good. After that, I applied to a similar project related to a military program. But at that time, there were many political issues related to it. Military projects are even more tightly connected to military and political issues. I did my best, but I couldn’t do that. Actually, this project was a subject that I really wanted to do, but I couldn’t. That was a severe memory for me.
Q29: So that touches on some things that you would have liked to have pursued but really haven’t had a chance to. And it sounds like the research funding—your funders haven’t had a strong influence on the direction you’ve taken. They haven’t really. They’ve allowed you to do what you’ve wanted to pursue.
Hyouk-Ryeol Choi: Generally speaking, it depends on the funding. Government-supported funding provides a very big freedom for research, mostly. But in the case of private company funding, I have to meet their requests. It is very much tighter. A long time ago, I did a project with a big company. Maybe the company name is known to you. At that time, I had to have a meeting every week on Monday morning at 10:00 with the person in charge of managing my project. Almost two hours every Monday morning. I had to provide research documents and discuss the results. It was really hard work. But it deserves it because I know what the purpose was. We had to deal with the private company, and that was a really good memory for me.
Q30: What would you say you view as your most important contributions to robotics?
Hyouk-Ryeol Choi: Well, as a robotic researcher, I think I am one of the first generations of robotic researchers in Korea, not in the world. Please believe that. I mean that I was among the first officially educated robotic researchers in Korea. Before me, there was no well-organized robotics program in Korea. So I think I was one of the first people in that area. After that, I educated almost 150 graduate students from my lab. I also started the concept of soft robotics in Korea in terms of research. And I continued my work on professional service robots for a very long time. There was a good chance to interact with company people while doing that. I think I provided very good ideas to companies. And finally, my work on physical contact issues—this is something that most people would like to have in robotics. I think I provided a really necessary technology for robotics.
Q31: How do you feel robotics has changed as a field over the course of your career? Looking back and looking forward, where do you see it going in the next 10 or 20 years? Any important developments that you think are coming?
Hyouk-Ryeol Choi: Well, it just came to my mind that robotics will be a major field of industry in the future. Now it is one of the hard issues, but a long time ago we repeated that robots are important. We said, “We have to do that. You can have one robot in the home, in the family.” But that was not the real reality at that time. Recently, I think it will come true. It will come true. So we have to be prepared for adapting to such a new world and changing world.
Q32: For that means, we have to develop related technology, and we have to change our style of life.
Hyouk-Ryeol Choi: We have to change the building methods of apartments. I mean that we have to make new ways for robots to move inside buildings and apartments for delivery services. Elevators will need to change, or we have to remove steps. That will be necessary for enjoying our life and living together with robots. Also, in terms of industry, I think we have a very nice opportunity to make money in robotics. Not just for intelligence, but in other areas—for example, very simple services using force and torque sensors. We can find many, many applications. It looks very simple, but I found that it produces a lot of value. We have to be ready for that.
Q33: So we’re going to shift gears a little bit here and talk more specifically about humanoid robots. You’ve done work in the area of humanoid robots or connected to humanoid robots.
The first question I want to ask you is: What makes a robot a humanoid robot? How would you define that?
Hyouk-Ryeol Choi: Well, finally, I think manipulation technology like a human is important. Humans can manipulate objects very softly. In doing that, humans transfer their feelings and emotions. So I think how to transfer emotion or feeling to objects and to people is important. A humanoid can do that. Well, I think this is the end of the world.
Q34: Can you say a little more about why you think that would be the end? Did you say the end of the world?
Hyouk-Ryeol Choi: I mean no more things to study or what? It is not a negative word.
Q35: Okay. So manipulation, so hands. But you know, you could imagine that someone would say, “Well then that’s not necessarily a humanoid, that’s a manipulator or that could be an arm.”
Do you feel like for a robot to be a humanoid robot, does it need to have legs? Does it need to have a head and a face? Could it just be a kind of trunk with arms and sensors that handles things in this soft or deliberate way that you were describing?
Hyouk-Ryeol Choi: I mean, having legs or having a head—that’s not the important thing. The thing is, the humanoid itself should be friendly to people, friendly to humans. That assessment comes from humans themselves. I mean that when humans meet the humanoid, they should think:
“Oh yeah, this is a really friendly creature.”
Then I think we have a well-done humanoid. For example, WALL-E is like a humanoid robot, and people think, “Oh yeah, it’s a really friendly one.” In that case, I think that is a very well-designed humanoid. In addition to actual operational performance, the robot should produce something human-like. Many things are important, but touching like a human softly and transferring some emotion through manipulation technology are very important. Whoever achieves such a minor transfer of emotion through physical motion—that will be important.
Q36: Maybe along those lines, can you share a little bit about why you’ve been interested in pursuing work on humanoids specifically, as opposed to other kinds of robots?
Hyouk-Ryeol Choi: I started my research career studying robot hands a long time ago. At that time, I studied robot hands without considering that humanoids were important. I just wanted to build a human-like hand and perform manipulation. But recently, many people say that a human-like robot hand is an indispensable subject for humanoids. So implementing feelings and manipulation skills into a human-like robot hand is very important for achieving a human-like humanoid robot.
Q37: Can I ask what you see as the social significance of humanoid robots? Ways that they are important for, what it is about them that is meaningful, socially speaking, particularly, I guess what I’m asking about is how do you see them existing socially? How are they going to be useful? What is important about a humanoid robot in a more social context?
Hyouk-Ryeol Choi: So the existing humanoid robot does not meet such kind of requirement yet, so it looks awkward and then hard surface, and then motion is not yet sufficient for meeting such kind of needs. But how can I say that? Just human-friendly humanoid that comes out to us. I mean, that can be very much utilized around us. Like domestic contexts or labor contexts or all of the above. Yes. In the first stage, I mean, it cannot help us with labor saving, but just standing and then servicing without any physical contact, that might be one of the starting points for a humanoid robot.
Q38: What would you say the work that you’ve done was first relevant to humanoids? Would you say that from the beginning, since you were working on hands, it was always relevant, or is there a point, a moment in your research career, that you would point to as “here’s where I really began working on something explicitly relevant to humanoid robots”?
Hyouk-Ryeol Choi: So I didn’t start the robot hand research considering humanoid robots, but eventually it became very closely related to humanoid robots. Instead of that, I also worked on quadruped working robots and legged robots. The legged robot—I didn’t expect it—but we received a lot of requests from industry. Many people said that they need some extraordinary mobility robot that can move inside a not-well-arranged environment. In addition to that, if they have some very minor equipment for doing their mission, for example, measuring temperature, gas, or other conditions, then extraordinary mobility and sensing technology are enough for them. So I did research on quadrupeds, and then many people started researching humanoids recently. I started the biped walking robot, but I don’t have a clear view yet of why a biped walking robot is necessary in reality. But anyway, some extraordinary mobility will be needed in the future and also at the current stage. That is my feeling.
Q39: I’d like to ask a little bit about the relationship between a robot’s body and its controller. How do you view that relationship? Is there anything you could expand on there?
Hyouk-Ryeol Choi: Body and control? Yeah. Well, it’s closely related. Previously, people studied model-based control systems, but such kinds of controllers are not prevalent recently. Most people are working on learning-based controllers. So even though learning-based controllers are good, we still need very good hardware, a good body, and a good kinematic structure and linkage structure. They are very closely and tightly related in terms of control.
Q40: Do you have any memories of any particularly interesting anecdotes or experiences? These could be unexpected successes, unexpected accidents, or failures in the course of your research over the years. And if any of these are any kind of funny anecdote or surprising anecdote, would you be interested in sharing?
Hyouk-Ryeol Choi: Let’s see. Well, I don’t have any special memories on that, but some difficulties always happen.
Q41: Okay, so you’ve talked a fair amount about application of research that you’ve done through your collaborations with industry, through your own startup. Is there anything else about that that you’d like to share, and maybe specifically any effects that you’ve seen from your research being applied in real-world situations that you’d want to talk about? And also, anything that’s been surprising to you in terms of seeing your research out in the world?
Hyouk-Ryeol Choi: Oh, what is that? So, I’m sorry, is it in the list of your questions?
Q42: This question is in the traditional list of questions that we’ve had.
Hyouk-Ryeol Choi: Yeah, yeah. So unfortunately, I don’t have anything special.
Q43: That’s okay. If you happen to think of something, you can share it.
Hyouk-Ryeol Choi: Yeah, yeah. There are only a few. If I have some special one, then I can tell you. But, well, minor things happen. You did talk about the teeth sensor, the dental system, which is pretty interesting.
Q44: Yeah, that can be.
Hyouk-Ryeol Choi: Yeah. Well, all of a sudden, a professor in the United States, in the dentistry department, contacted me because he would like to build this kind of setup. Actually, at that time, he spent almost four times the expensive price for building such kind of experimental equipment. But we told him that we could make it at a quarter of the expense, including all the software. So we delivered it to him. He is a very young professor. What the young professor does is write many papers. In his paper, he wrote one sentence that this equipment was produced by AIDEN Robotics with the address of my company. After many decades, I found that this paper had a very strong force for spreading the reputation of my skill and research technique to people. That was a very memorable and good moment for me.
Q45: Yeah, it’s pretty good. Very good advertising inside of a research paper.
Hyouk-Ryeol Choi: Yeah, yeah, yeah. Even Korean dentists contacted me. Many, many.
Q46: Good one. So one of the last questions we like to ask is if you have any recommendations or advice for students who are interested in getting into robotics.
Hyouk-Ryeol Choi: Well, robotics is a very hard-working field of area. So if you do not have a really lifelong interest in this robot area, you cannot do that. But I think if you have some success in robotics, then it is very deserving to work on it. I think so.
Q47: And the last question we would like to ask is if there is anything that we should have asked that we haven’t thought to ask, or anything that you’d like to bring up or share that we haven’t talked about yet.
Hyouk-Ryeol Choi: What is that? You always ask me difficult questions. I’m sorry. Well, a long time ago, I took a short movie from the Korean Broadcasting station, and they still spread the movie through the YouTube channel. At that time, I talked about the human-friendly robot. Everything that I have done is related to something human-friendly. To make the robot friendly to humans, to help humans, and to be friendly to humans. By doing that, I feel comfortable, I feel achievement, and I feel happy. So anyway, we can do many, many things: software, hardware, humanoids, hands, whatever. But we have to keep in mind that what we are doing should be helpful to humans and should make humans happy and friendly.
Q48: Anything else you want to share?
Hyouk-Ryeol Choi: Nothing special.
Q49: Well then, I want to thank you for sharing your time and your experience with us.
Hyouk-Ryeol Choi: Thank you. I couldn’t respond to all of your questions nicely.