Interview Transcript
Q1: To begin, could you tell us your full name, date of birth, and where you were born?
Jingang Yi: My name is Jingang Yi, and I was born in China. I did my undergraduate and Master’s degrees back in China, and then I moved to the United States for my Ph.D. After that, I stayed in the U.S. for my career.
Q2: Would you like to tell us a bit about your childhood and what your parents did?
Jingang Yi: I had a happy childhood. I grew up in a smaller city in the central part of China. Academically, I did pretty well. I was not the oldest child; I have two older brothers and one younger brother, so it was a big family—quite typical for China back at that time.
I really enjoyed studying engineering because I was good at math and STEM subjects. High school and getting into college was very hard and competitive. I did my Bachelor’s in Electrical Engineering at Zhejiang University in Hangzhou, China. After my Bachelor’s, I did my Master’s in Mechanical Engineering at Tsinghua University in Beijing. So I had a very interdisciplinary educational foundation combining mechanical and electrical engineering, which my career was built on.
Q3: And then you moved to the U.S.?
Jingang Yi: Yes, I came to the U.S. in 1996. I went to UC Berkeley in California for Mechanical Engineering and did my Ph.D. there. Along the way to getting my Ph.D., I was also very interested in mathematics, so I completed a Master’s in Math as well. California not only had nice weather, but it really gave me the training my career is built on.
As I was finishing my Ph.D. in vehicle dynamics and control—I was trained as a control engineer—I started my career in industry in 2002. I joined a company in Silicon Valley working on semiconductor manufacturing for about two and a half years, which was a dramatic change from my Ph.D. research.
Q4: How did you get into robotics in the first place?
Jingang Yi: That’s interesting. In that company, the main focus was semiconductor manufacturing, so I started with manufacturing automation as a trained control engineer. Back then, robotics wasn’t emerging the way it is right now; it was still a traditional engineering discipline.
My journey to robotics started with manufacturing automation, and since I worked on vehicle dynamics and control, I also tried working on mobile robots. When I tried to publish my results, I aimed for international automation conferences, and IEEE ICRA (International Conference on Robotics and Automation) was one of the main venues. I still remember attending my first ICRA in 2005, where I had my first ICRA paper.
Q5: Going back further, growing up in China, did you have popular cultural influences or a particular image of robotics in your head?
Jingang Yi: In terms of popular culture, mechatronics was the key buzzword at the time. I studied mechatronics engineering during my undergraduate years, and I have very strong memories of trying to learn and apply it.
Q6: What would you say was the very first major program or direction?
Jingang Yi: I was recruited to Zhejiang University because of that major. It was a new major, similar to how robotics programs are starting up in many U.S. schools now. When I was an undergraduate, mechatronics was one of the emerging areas. People believed that by joining mechanical, electrical, computer science, and control engineering, mechatronics was the future.
Q7: What were some of the main research questions you were trying to tackle when you started out?
Jingang Yi: One of my first research topics was: how do we control a car just like race car drivers do? That was something that really fascinated me. I was trying to figure out how we could drive a car like professional drivers, which was closely related to my Ph.D. research. People are still working on that problem today using different techniques, but that was my first interesting research question.
Q8: How has your primary research question changed over time throughout your career?
Jingang Yi: Research interests are always dynamic. As a professor and researcher, interests evolve over time, though tackling research problems takes significant effort over many years.
I started with mobile robots and race car driver behaviors to understand how to control a car in unstable regions—trading stability for agility and high performance—and then moved into controlling autonomous vehicles.
Over the years, I became interested in human-robot interaction and legged robots. Over the last ten years, I’ve worked extensively on legged robots, always searching for challenging problems and unusual solutions. For example, we studied how humans react to trips and slips in unstable behaviors—similar to pushing cars to their stability envelopes.
Recently, we’ve been working on how legged robots navigate changing terrains like sand, mud, or loose ground, analyzing how environmental disturbances impact stability, locomotion, and efficiency. We then translate that understanding into building more resilient, stable, and efficient robots.
Q9: Did you encounter any major financial, institutional, or technical challenges in your career?
Jingang Yi: Going through the tenure process is always challenging for any academic. At early stages, getting funding, setting up the lab group, and recruiting students are major challenges.
At the same time, my wife is also a professor, so we had to navigate dual-career challenges alongside raising a family and kids. Different stages bring different challenges, but finding enough interesting problems to work on with your expertise while constantly learning new techniques makes being a professor very enjoyable.
Q10: Were there any tools or techniques that came up during your career that were major breakthroughs or opened up new possibilities?
Jingang Yi: I was really impressed by the growth of machine learning. When I was an undergraduate or graduate student, people tried to stay away from neural networks because they were viewed as unreliable “black boxes” that were difficult to generalize.
When the new machine learning wave came in over the last ten years, we initially had doubts thinking, “Oh, these old problems are back.” But over the years, I’ve been fully convinced that this is definitely different and transformative. Machine learning has changed our research, our work, and our daily lives in a very positive way.
Q11: Who have been the main funders of your research throughout your career?
Jingang Yi: Throughout most of my career, my research funding has come primarily from government agencies and foundations, mainly the National Science Foundation (NSF).
Q12: Is there anything else in your career that you wanted to pursue but couldn’t or didn’t get to do?
Jingang Yi: There are always good ideas, but you have limited time, energy, and resources. I wish I could have made more progress on certain research problems, but due to limits on student capacity and experimental setups, you always feel you could do better. My expectations are always higher than what I’ve achieved, but if I had more time to work on some of those problems, they probably would have turned out even better.
Q13: How do you think robotics conferences and the broader community have changed over the years?
Jingang Yi: When I started attending ICRA around 2005, there were probably around 1,000 attendees. This year, I heard there are nearly 9,000! So I’ve seen dramatic, rapid growth in the community over the last five to ten years. Machine learning has been a major driver, but new emerging hardware technologies—like advanced sensors and actuators—are also pushing the field forward. It’s a totally exciting time to work in robotics right now.
Q14: Where do you think robotics is going to be in the next 10, 12, or 20 years?
Jingang Yi: I think robotics is going to change our daily lives. We are going to see more robots moving out of research labs and into daily life—at home and in factories. Certain jobs will definitely be impacted or replaced by robots as they work alongside humans. While it might not happen as seamlessly as some describe right away, a significant amount of work in manufacturing, industry, and daily life will be handled by robots.
Q15: Have recent developments in generative AI made a significant impact on robotics?
Jingang Yi: Yes, it has definitely had a big impact, though I think the transformation in robotics will be slower compared to fields like vision or language processing. Robots are physical systems, and due to the high dimensionality of physical interactions and environmental complexity, generative AI’s impact on robotics will be more gradual—but it will happen.
Q16: Could you share an example of your research being applied to real-world situations and societal impacts?
Jingang Yi: About 10 to 15 years ago, I worked on infrastructure robotics funded by NIST and the U.S. government. We focused on developing robotics for infrastructure inspection. Most U.S. bridges and highways were built over 50 years ago, and maintaining aging civil infrastructure is extremely costly and difficult given the scale.
We developed a robotic system integrated with non-destructive infrastructure inspection sensors to safely, efficiently, and cost-effectively perform maintenance and inspections. Startup companies were eventually built on this technology, so it’s exciting to see our research translated into real-world commercial and societal impact.
Q17: In terms of real-world developments, has anything surprised you recently?
Jingang Yi: I’ve been really surprised by how fast humanoid robots have developed in the last couple of years. Seeing how rapidly humanoid robot skills are growing—enabled by machine learning techniques—has been truly surprising to me.
Q18: A lot of your work is with the NSF, but have you collaborated with industry as well?
Jingang Yi: University-industry collaborations are not always easy because academic focus tends to be more on the fundamental side. We have collaborated on projects with companies like Siemens and Ford, but aligning common interests and overlapping focuses requires effort, so it doesn’t happen for every research topic.
Q19: What would you recommend to students who are interested in getting into your field today?
Jingang Yi: It is a great time for students to get into robotics and machine learning. My suggestion is to be open, talk to people, and attend conferences like ICRA. From my own experience, I know it can be intimidating or hesitant to approach people at meetings, but every time I do, I gain a lot. You might find someone whose work is directly relevant, or get inspired by conversations with others. Be open, be ambitious, and always try to push your own limits.