Interview Transcript
Q1: Could you start by sharing your full name, date of birth, and where you were born?
Hyungpil Moon: I’m Hyungpil Moon, and I was born in 1973, June 16th in Seoul. And I still live in Seoul.
Q2: Could you tell me about your parents and your family upbringing? What was it like?
Hyungpil Moon: My father was raised by a single mom long time ago under the Japanese colony era. My grandfather passed away when he was very young, so my grandmother had to figure out how to raise children. They sold all their lands and farms, moved to Seoul, and opened housing in front of Yonsei University. My grandmother always asked my father to study hard, so study and hard work were always emphasized.
I am the fourth child and have three older brothers. In Korea, the entrance exam to university is very competitive. When my eldest brother was preparing for the exam, because of the lack of sleep, he hid somewhere in the closet just to sleep—that’s how competitive it was. Growing up watching that, I understood what I should do. Going to school and studying a lot was very natural to me.
Q3: Where did you go to school?
Hyungpil Moon: Luckily enough, I went to a prestigious high school where we were the first class of students. Everything was under construction—we even used the old middle school building that my father used to go to! Then we got the new building, so it was crazy with construction, but fun in some sense.
For my undergraduate, I went to POSTECH in Pohang. It’s a small private engineering school. Since everybody lived in the dormitory—which is rare in Korea because most students commute—every night was party night and we had a great time together.
After my Master’s degree, I finished two years of mandatory military service and then went to the University of Michigan, Ann Arbor in 2000 for my Ph.D.
Q4: Who did you work with at the University of Michigan, and what was it like?
Hyungpil Moon: My advisor was John Edwards, who was an assistant professor back then, and I was his first Ph.D. student. It was very tough because he was starting his career as a junior faculty member. Michigan had a great program in manufacturing and controls, and he was very busy, so meeting with him was quite challenging.
Q5: Did you go work outside after your Ph.D.?
Hyungpil Moon: After my Ph.D., I went to Carnegie Mellon University in Pittsburgh at the Robotics Institute with Howie Choset for almost two years as a postdoc. There, I learned a lot about new robotics areas like SLAM (Simultaneous Localization and Mapping), which was very hot back then. We also formed a study group to study geometric aspects of robot control with peers like Alberto Rodriguez (who is now at MIT/Boston Dynamics).
Q6: How did you get interested in robotics in the first place?
Hyungpil Moon: Robotics wasn’t very popular when I was an undergraduate. I was originally interested in thermodynamics and fluid mechanics because they were mathematically and conceptually interesting. But during an internship in a robotics lab, I was tasked with computing the moment of inertia for a new robot design that a Ph.D. student was developing. I didn’t use a computer; I hand-drew and measured everything and calculated it by hand. My senior student was so happy because my calculation was so accurate. That made me realize how useful these mathematical computations were in robotics, so I decided to go to graduate school for robotics.
Q7: What was the first robot you heard about or experienced?
Hyungpil Moon: As a kid in Korea, I watched cartoon movies like Japan’s Astro Boy and Korea’s Robot Taekwon V. But in real life, it was during my undergraduate internship when I saw the PUMA 560—a 6-degree-of-freedom manipulator. That was the first manipulator robot I experienced.
Q8: What were some of the main research questions you had when you started in robotics?
Hyungpil Moon: When I started my Ph.D., I wanted to use simple actuation to perform complex tasks on multiple objects—like luggage sorting at airports using conveyor arrays rather than picking up each box with a complex hand. This non-prehensile manipulation (manipulating objects without grasping) was my early research focus.
Later, when I became a professor in 2008, I received Korean government funding to develop robotic hands to achieve dexterous manipulation similar to what human hands can do.
Q9: How is your research funded, and does the funding source change how you conduct research?
Hyungpil Moon: About 90% of my funding comes from the Korean government, which supports more fundamental science and technology development (up to TRL 3–4). Recently, private manufacturing companies have also been funding us because they are very interested in applying new Vision-Language-Action (VLA) models to their manufacturing processes.
Q10: What are the biggest challenges you are facing right now?
Hyungpil Moon: The rapid development of AI technology. Two years ago, GPT wasn’t perfect at my exam problems. Last year, it got very good. This year, it solves textbook engineering problems step-by-step and provides explanations better than I do! As a mechanical engineering and robotics professor, I constantly ask myself what I should teach students and how we can adapt to this new era rather than sticking to old ways.
On the research side, collecting data for physical manipulation is extremely difficult. For instance, to train models on part mating, teleoperating a robot to put parts together was too hard for operators. So instead, we had operators take parts apart, reversed the video data, and used that to train the model.
Q11: Are there any real-world applications of your research that you’d like to highlight?
Hyungpil Moon: We recently developed a truck unloading solution for logistics hub terminals. In Korea, trucks arrive overnight filled with about 2,000 mixed parcels, and humans have to unload them onto conveyor belts within 50 to 60 minutes. It’s grueling physical labor.
We developed a special sliding-surface gripper that drags boxes together onto a low-drop basket conveyor rather than picking them up one by one. This solution allowed us to unload 2,000 boxes per hour safely, handling various shapes, sizes, and unknown weights.
Q12: What would you recommend to students who are interested in going into robotics today?
Hyungpil Moon: First, face the problem directly by working with real physical systems and real data, even if it’s a small mobile robot. Second, participate in robotics competitions organized by professionals. Experiencing close-to-real situations helps you identify where the actual bottlenecks are—whether it’s sensing, hardware, or algorithms—so you know what mathematical and programming skills you need to develop.
Q13: Is there anything else you would like to add or discuss?
Hyungpil Moon: I strongly believe in the value of robotics competitions as standardized benchmark testbeds and drivers of technology direction. For example, starting around 2015 at ICRA/IROS, I organized the Indoor Autonomous Drone Racing competition for several years. We watched vision-based autonomous algorithms evolve from scratch to eventually matching human pilots. I hope to continue promoting these kinds of benchmark competitions in the robotics community.