Interview with Jun Ho Oh

Q1: Hey, professor. Thank you so much for coming today. Could you tell us when did you first begin walking in humanoid robot?
Jun Ho Oh: It was. Year 22, right after bicentennial. Okay. Yeah, I started I was motivated by Honda’s Asimo, and it chucked me.

Q2: What problems were you trying to solve at the time when you started working at humanoids?
Jun Ho Oh: By the time I didn’t have any information and I didn’t have any idea of what is humanly self. By the time I thought that it would be impossible to make human like machine like Asimo, because before 2000 there are some research related with bipedal walking, but it’s just just pure research purpose. So it looks a little bit weird and just imitating the baby walking, but it’s not very serious as far as as far as I know by the time. So first thing that I’m thinking is that, what kind of actuator should be used and what kind of sensors and what is the what capacity will be fine for the computer and batteries? Nothing. So I don’t have any idea. So I start from the scratch. What I had by the time is that I have had many experiences dealing with motors and microprocessors and computers and real time control sensors. So I believe that if I combine my knowledge with those kind of mechatronic stuffs, I couldn’t make it. That was your…

Q3: Is there any publication that we should check out?
Jun Ho Oh: Yeah. In the early stage I forgot the full name of the paper, but we present that one in year 2004 or 5. That was one of the earliest paper. Yeah. Yeah.

Q4: And how did your work evolve over the years?
Jun Ho Oh: At first we didn’t have any concrete font, so I used very small money collected from my colleagues. By the time and some basic research money when I release second version of humanoid. In year 2003, I started to get from fund from the government, and then I continued my work for more than ten years under the government support.

Q5: Okay. And what are the technical challenges have you been trying to tackle over the years?
Jun Ho Oh: Yeah, most of them are from the actuated. So we need strong and light and educators and control driver boards type thing. Actually two related things. And secondly, sensors, we should have very nice inertia sensors and must be very small. But very low drift is very important. Yeah. Most people thing is the working algorithm itself. Okay. We have to make it work. This was hardest one. Yeah.

Q6: Can you tell us the key publications that we should take from the past three, almost three decades?
Jun Ho Oh: We intensively wrote a paper at first stages after. There are many small details. So major one is first three papers, I guess. Yeah, yeah.

Q7: Do you have the titles?
Jun Ho Oh: I forgot that, yeah.

Q8: And it’s a bit of a question about software and hardware relationship. So how do you see the relationship between a robust body and its controller?

Jun Ho Oh: Okay. There are two kinds of controllers inside. First of all is control itself should have some sort of level of control. It must be very robust. And it takes all the comments from the main computer. Anyhow, that actually itself is computer itself. And secondly there is a motion controller. It generates all the motion and control the robot to be stabilized in all the way and make it work. Those are two computers. If I add one more computer is kind of related with communication and kind of intelligent and giving the major command walking directions or something like that. So taking care of scheduling or something. So we use one computer sometimes two. But yeah at the first stage we use only one. But later when you put the vision related things, then we added one more computer to dedicated for the for those purposes.

Q9: Is there any specific approach that you take design humanoid robots?
Jun Ho Oh: Not really. Only thing is that we have to make everything be compact and lighter and also robust. I’m thinking all the way if when we want to have some kind of external case, eastern case, then we have to be very careful that some kind of small not to extrude outside. So everything must be confined in, in, in the restricted case. This is another concern. Next most important one of the important thing is that heat problems thermal problems because motor and computer generates hits all the way. If it drives, the robot be more excessive power. The general is big hit. So we have to find a way how to remove the heat. So sometimes we work through the air cooled and the passive cooling. Yeah we used many different kinds of cooling method major.

Q10: So starting from Asimo and the moment you got inspired during your career, you witnessed a part of the evolution of design and control of humanoids, and also the first steps into the commercialization of humanoid robots. Could you give us a brief account of this historical development that you witnessed?
Jun Ho Oh: Yeah. Yep. In Korea, we at first we are the only research group working with humanoids. At first, a couple of years, and then the c’est c’est. They started work 1 or 1 year later. Then I started… And also… At first the sensor electronics tried to have their own humanoid at first, but they, they, they released the results in 2012. Somewhere in a song I cross our I and they stopped the research on humanoid the. After releasing the final results and some universities Chi and snu and and they continue to work on humanoids and these days there are 2 or 3 I don’t know so many maybe five human companies started. They are traditionally dealing with robot kind of kobolds or service robots, etc. they also actually the components. They now add humanoids in their shelf.

Q11: The first challenges in these early days, the humanoid robots. Well, what was that? What were the challenges like first challenges that community was trying to tackle in these early days?
Jun Ho Oh: Lack of information and human to make human to walk or build a humanoid. You can we can refer all the papers, recent papers, but it does not say in 100% the main primitive work. About 80% of work is in behind by experience and no house and try and errors. It is not released on the paper, it cannot be appeared on the paper. So you have to take that kind of knowledge and experience by himself. It takes time. That is quite hard.

Q12: Okay. And with recent emergence in AI and in these things, how does it affect humanoid robotics?
Jun Ho Oh: A big as I mentioned yesterday in my talk, that we tried humanoid to be used in your unstructured life very many, very many ways. And actually honestly, it failed actually. So we proved that, yeah, it can do that, but it’s not quite realizable in your life. So we stopped and yeah, let’s wait for the next technology coming. So one day, one day AI appeared. So it’s the second shock when I started after studying The Human, of course, I did some research on quadruped. We accepted some AI technology inside, but it cannot be realized. I don’t believe that it can be. It can be realized in the robot. But there is some one Japanese. The Chinese company unitary released that one just on one half years ago, not a long time. And so, so high speed. So I changed my kind of concept for the humanoid that my preoccupied concept. The idea is that, you know, monoid will appear in my home until in my life. Okay, time. But now I’m thinking that it might be possible. Okay. It’s coming. So I’m thinking in a different way. Start to think. I’m not very optimistic, but I’m not a pessimist. Now. Just live a conservative. Plus. Exciting. Yeah, exciting.

Q13: Do you remember any unexpected success or failure?
Jun Ho Oh: Yeah. Okay. You asked me before that. How could I commercialize the robot? Okay, I’ll tell the story. It will be more here at this time. Year 2012. I remember 11 or 12. I got some kind of request to deliver. Human robot, my human robot in the United States from the several colleges. At the same time, it’s the one program in the United States. They asked me to deliver six humanoid human robots to the United States for each universities. And also almost at the same time, I have two orders from the Singapore National Research Institution. So we are asked to build eight robot, but it’s impossible to handle that kind of things with two students. So I had to kind of set up a kind of company enterprise for to take care of that, those kind of requests. So yeah, I founded Rainbow Robotics by the time and first one is we just we first got older and make company and then produce and sell. Okay. So it was kind of very in financial way. We didn’t have any difficulties to because we we can’t afford it at first. And they started business. So at first the order is about our 4 million USD, quite big by the time. And there was a start of commercial robot.

Q14: Is there anything that was a challenge?
Jun Ho Oh: Okay, so at first I didn’t have any idea about to participate in the upper body challenge, but I was called from the university who bought my six robots in the United States. They say they’re going to participate rubber challenge. So they asked me that, could you modify or build the robot to take care of that kind of challenges? Say, yes, yes, I said yes, and I redesign and that one and sent to the United States. But the thing is that there are six different universities in different locations. So kind of collaborative work is not quite silky.So I worried a lot and we give them an instruction how to dealing with that one, but even hard to communicate with them. So I decided okay, let’s try our own. So we made team Kaist with the similar robot. So I, I played in two different places into two teams anyhow. At last, as a result, we won the trophy. The game challenge.

Q15: This second kind of events for me and third one is.
Jun Ho Oh: The business opening its IPO to the market. And we were very sensitively open our my company in an IPO in the stock markets. And value was quite reasonable by the time. So it was very good for us and we got the fund from from those events. And fourth one is we were company was named by Samsung Electronics after two years later to the IPO. So yeah, there’s a story of event in the and some part of it.

Q16: Fascinating success story. Thank you. So could you you mentioned Asimo from Japan, but could you pinpoint maybe labs or people that were influential in human or robots go about the past 50 years? But what do you say? Can you name people or the labs have been influential in humanoid robotics?
Jun Ho Oh: Yeah, of course there are so many, so many.But if I ask you like that way, my one of my viewpoint is I don’t remember. I can remember names, places and time and list. I, I have many acquaintances in this area. I know very many people, but unfortunately, I couldn’t remember this. This is my brain problem, so I’m sorry I cannot list. Yeah, yeah. In Germany, in the United States, in Switzerland, in Italy, Japan. Of course not. Yeah, yeah, I know, but also on dynamics of course. Yeah. So those are so popular. So I just imprinted the name and time of the title of the company is imprinted my brain, not others. They are all famous and, but I’m just popping up the name. But I couldn’t say the order of.

Q17: Reference you like to share with us? Please Send us later. To wrap this up, I’d like to ask you one last question. What are you most excited about as this humanoid conference today?
Jun Ho Oh: Yeah. As I mentioned earlier, that we are kind of dark ages since since last ten years. And it and things that when AI comes to the human human society, traditional researchers was very embarrassed because they are using a couple of different technologies and looks very different. And it might be realized then the in my realized that we thought it was unreal. So, this year. So I didn’t participate human conference last five years. So this I this time with call and human conference at the same time. So it’s very different and so many humanoid robot manufacturer not only in China, in Korea, in the United States and other European countries. It was quite amazing for me to see the kind of things that same places, same place, their quality is very high and somehow somehow normalized, so generally quite become higher and higher. And now it’s time to start to mix up with traditional the robot concept, plus new, new coming AI based robot technology. It’s kind of merge time, so I’m expecting that coming couple of years. We will see quite dramatic improvement or development or progress in human, I hope, and I want you to be a part of that.

Interviewer: That’s very exciting to end this interview. Thank you so much for your time.
Jun Ho Oh: Yeah, I enjoyed a lot. Thank you