CSpotlight: Cleaning Up the Ocean One Robot at a Time

Why did you choose to pursue a degree in computer science specifically at the University of Minnesota?

Around the time I completed my master’s degree, I was looking for schools where I could pursue my research area in robotics. The University of Minnesota has a strong robotics program, and we have professors who specialize in many different robotic fields. Examples include underwater, manipulation, agriculture, drones, etc. That's why I chose computer science at the University of Minnesota. Also, I expected that computer vision and machine learning research at the UMN would allow me to develop my perspectives in various research fields.


How did you become interested in computer science? What are your specific interests within the field?

My bachelor’s and master’s degrees are in electrical engineering. During my master's program, I took a course on machine learning. That subject inspired my general interest in artificial intelligence, which led me to more in the robotics field. Robotics is a broad field that involves computer science, electrical engineering, mechanical engineering, you just name it. When people think about robots, people often think of humanoids that are walking and moving like a person. But my interest in robotics was how to make robots understand their environment using robot vision and machine learning. That is most relevant in computer science, and that's why I chose computer science over electrical or mechanical engineering. I am interested in improving robotic perception and deploying robots in potentially dangerous environments (for example, underwater, space, and disaster scenes) instead of risking human safety or lives. As much as I am passionate about developing robotics algorithms, I am interested in deploying a robot with my algorithms in real-world environments.


Congratulations on being a research assistant! How will this role impact your academic and extracurricular work?

I’m a research assistant and was supported by a MnDrive fellowship too. Being a research assistant gave me the freedom to invest most of my time in research which is my passion. When I had just started my Ph.D. program, my research agenda was broad, which was robot vision. With the research assistantship and fellowship during the early years of my graduate studies, I could explore different branches of research topics in robotics and refine my research topic within Robotics. 


Tell us more about your internship experiences.

Recently, I completed a 7-month-long research internship at Samsung AI (Artificial intelligence) Center in Manhattan, New York. It gave me tremendous inspiration for working in the real field with talented researchers and mentors in the group. It was undoubtedly a great opportunity during my Ph.D. program to discover new perspectives on my research. During the internship, I worked on manipulation robots. These robots were designed for indoor environments like at home or at an office. The expertise in robotic visual perception gained from my research experience at the UMN gave me a strong foundation to address research problems from my internship. Setting up clear deadlines during the internship helped me to focus on research. As a result, I published two papers at top-tier robotic conferences (ICRA and IROS) from my internship projects.

Another internship experience I had was with Sentera, an agricultural drone company based in Minneapolis. During the internship, I gained research experience in the agricultural domain, and I worked on developing a weed detection algorithm in fields. With the algorithm, the drones fly over crops, collect data, and analyze it, so farmers can make smart decisions based on the detection result. From these experiences, I expanded my research agenda beyond the underwater domain.


What do you hope to contribute to the computer science community at the University?

I hope my research inspires other students in the Department of Computer Science & Engineering to be involved in robotics research and develop their own research topics. For me, it was not easy to catch up with my computer science background as I changed my major from electrical engineering, but I want to show them hope and possibilities. 


Have you been involved with any research on campus?

I work with Professor Junaed Sattar. My research lies at the intersection of field robotics and robot vision. When I say robot vision, it means how the robot understands things from the visual information it receives from a camera. I have developed algorithms for improving robotic vision in unstructured environments such as underwater and agricultural domains. 

I began my research with underwater robotics and the underwater debris clean-up problem. The underworld debris problem is a huge issue. There is a study showing that people consume about a credit card size amount of plastic every week because of the microplastics that exist in the water. That problem really intrigued me and inspired me to develop one of my research goals toward microplastic clean-up using robot vision. Particularly, I have worked on detecting objects in data-scarce and visually challenging environments to address this problem.

Additionally, I became interested in complementing robots' perceptual capabilities with human-robot collaboration. This is important where, as you can imagine, robots are not smart enough to do tasks independently. The idea was to make the robot utilize human knowledge to conduct tasks when it cannot understand its surroundings or discover objects. For instance, this can save divers from directly collecting debris, which poses a high risk of exposure to toxic materials. Instead, we can have robots find and remove debris by using human knowledge of debris.


What advice do you have for incoming computer science students?

For graduate students, I recommend having an open mind and talking to people outside of your research area. It'll give you a fresh perspective on the problems that you're working on. Oftentimes we work in our little worlds and miss the big picture. We need to step out of them to gain alternative points of view. For undergrad students, I would suggest they expose themselves to research experiences using either voluntary opportunities at the lab or over the summer. You can utilize the REU (Research Experiences for Undergraduates) site from the National Science Foundation. It will help narrow down your interest even if you don’t plan to go to graduate school. 


What are your plans after graduation?

My long-term research goal is to enable robots to perform tasks autonomously in unstructured environments with robust vision. I want to continue my research to achieve the goal in either academia or industry. Currently, I'm leaning towards academia because I want to develop my research agenda further and enjoy collaborating with other disciplines. I'm excited to continue and build on my research in the robot vision area. With improved robot vision, I believe we will see more robots around us and they will help save human lives from dangerous tasks. 


Are there any additional experiences you did that you would like to highlight?

I was a MnDRIVE scholar, where I taught K-12 students at the tech camp, which goes over the basics of engineering and robotics. I enjoyed that experience, and I would recommend other students look into that. Many students had little to no computer science experience going into that camp. It was an excellent opportunity to see how they get inspired by learning about new technology. Some students even wanted to pursue careers in robotics after the tech camp experience. That was an inspiring experience for me, and it was truly rewarding. That reminds me of the importance of outreach and community services besides working on my research.