Master's Profile: Yan Zhang

Yan ZhangCurrent Position: PhD student in Mechanical Engineering/Robotics at Duke University

Undergraduate: BS in Mechanical Engineering, Tsinghua University (China)

MEMS MS Program Path: Master of Science (MS) in Mechanical Engineering, 2016

Why did you choose Duke for your graduate study?

I heard great things about the faculty at Duke, and I also liked the fact that I can make very flexible personal study plans by taking courses from multiple disciplines as long as they can help my research. Additionally, Duke Engineering holds a lot of social and professional events for me to reach out to people and develop skills that would help me build my career. At Duke, I not only treat my research as my job and learn to be professional, but also treat it as part of my life and have fun with it.

How did your time as an Master of Science (MS) student at Duke prepare you for your PhD?

The Master's Program at Duke gave me plenty of opportunities to get actual experience in advanced research and helped me focus on what my research interests truly were. I knew I wanted to pursue a PhD when I came to Duke as a master's student, and while there were lots of topics I was interested in, I wasn't sure I was prepared enough to jump right into a PhD. The MS program provided me with a great foundation and after taking a few advanced courses, I got a feel of what kind of tools I would be using to conduct research and I grew more confident in my own capabilities. More importantly, I got to talk to a lot of fantastic professors and graduate students, and we had the chance to exchange opinions about various research interests and their experience both in academia and industry. All of these helped me form my interests and plan for my own career.

What was the most valuable part of your Duke experience?

My interactions with enthusiastic professors and graduate students was a valuable part of my Duke experience. Talking to these great researchers let me learn more than what you would typically learn in published papers and these interactions helped me form my own philosophy about research and topics I want to pursue.

What were the most useful classes you took at Duke?

While all lectures I took here were well-taught and very informative, the most informative were Linear System Theory, taught by Michael Zavlanos, and Engineering Thermodynamics, taught by Adrian Bejan. Linear System Theory is essentially what got me started on my current research. The content of this course is fundamental to other control topics, and can also be related to other useful research topics like numerical analysis. Although the content in Engineering Thermodynamics isn't closely related to my own research, I was still impressed by the lectures from Dr. Bejan. Not only did he teach us about the theorems we needed to know, but he also taught us how people built the field to solve problems. I realized conducting research is not just about caring about your own small field, but also getting to know everything about your communities, their history, their initial motivations and all the success and failures they made along on the way.

What advice would you give to someone considering a master's degree in MEMS at Duke?

Be self-motivated and reach out. There are plenty of resources offered at Duke. The faculty and the graduate students are nice to talk to and they always welcome you to engage in their research. They give good advice, but you cannot and should not expect them to make decisions for you. To really succeed, you cannot just stay at home, take lectures and finish homework. For students who plan to keep on a PhD track after their master's degree, they must read papers and engage into actual research as much as they can. Reach out to faculty here and find work to do. No matter whether the work is important or minor, you can always learn more than what you are assigned to do.

What is your current research about?

I'm working with Dr. Zavlanos for my PhD, and we design algorithms for critical robotics issues. For example, we developed algorithms based on optimization methods or dynamic programming to let robots with on-board sensors actively make decisions on where to take the next measurements to get information from their surrounding environment in a more efficient way. We are also interested in developing algorithms for multi-agent systems so that multiple robots can coordinate to accomplish some tasks – sensing, routing and learning, especially when the environment is uncertain.

My current research involvestrying to characterize the performance of the distributed controller or information filters of a multi-agent system when the agents have very limited computation and communication accuracy. In my research, I acquire and apply knowledge in robotics, mathematics, statistics and computer science to develop new theoretical results. I also work on hardware and do some coding to implement experiments to validate our proved results.