On Teaching Robotics
[ personal ]

The following are some reflections that were requested from me by the School of Computer Science at Carnegie Mellon University after I was awarded the 2024 Alan J. Perlis Graduate Student Teaching Award. I am placing these here in the hope that these reflections may eventually be helpful to someone.

Teaching robotics is a unique opportunity. An effective lesson in robotics requires teaching theory and practice simultaneously. The theoretical elements are from several areas of mathematics and physics while the practical aspects require many concepts and tricks from engineering. I derive a lot of joy in operating at the intersection of theory and practice. Consequently, all activities like deciding learning objectives, designing lectures, creating assignments, delivering feedback, and helping students become enjoyable when I am teaching robotics. I doubt that I will be able to teach anything else with a similar commitment.

Teaching robotics requires instilling visual intuition. How is the robot moving if a particular force is applied in a different direction? Such questions require cultivating a sense of imagination and curiosity in the student. I found it fruitful to invest time in creating illustrations and animations towards meeting this requirement. Perhaps unsurprisingly, the tools I learnt in this process helped improve the quality of exposition in my research papers too!

Teaching robotics requires a good testing infrastructure. How do we tell a student that what they did to the robot is correct? How do we measure correctness and deliver feedback? How do we do this in an automated way? In computer science courses, automated evaluation and grading of code submissions has existed for a while. Extending this to robotics is an ongoing challenge because robotics involves hardware that moves in 2D and 3D space due to software. While I have not had the opportunity (yet!) to teach a course with both hardware and software components, it is going to be exciting to create infrastructure supporting such courses.

Teaching robotics in 2024 and beyond is important. The robotics industry is booming and will appreciate all the talent it can get. Aspiring engineers and researchers need to know past methodologies, the current state of the art, and the associated pros and cons. When we TA or instruct, it is important to keep in mind that students will probably be deploying robots in the real-world and impacting many lives. I found this fact to be motivating. It kept me going especially when I felt that teaching was taking up too much of my time out of research.

These thoughts on teaching robotics have developed through the opportunities I have gotten at the Robotics Institute. I am grateful to Prof. Wennie Tabib and Prof. George Kantor at the Robotics Institute for giving me the opportunity to teach through 16-362: Mobile Robot Algorithms Laboratory. Thanks also to Prof. Kris Kitani and Prof. Deva Ramanan for the teaching assistantship opportunities. I am happy that I was able to contribute meaningfully for the students, and in the process correct my understanding of several areas in robotics.