Together with students, I designed an educational robotics platform - originally called KuBots, now Plato Pod - where students learn programming by writing code that controls physical robots on a shared arena. The pedagogical idea is simple: when your code has a bug, the robot does the wrong thing in front of you, and no amount of copy-pasting from an LLM fixes that. Students have to understand the logic themselves.

The first-generation platform was co-designed and built with students at KoreaTech in 2014-2016. Students in class programmed differential-drive robots to perform line following, maze navigation, and multi-robot coordination exercises mapped to weekly lecture topics (sequential execution, conditionals, loops, functions). A student survey (n=22) showed that 95% rated the lab experience as "Good" or "Very Good," 77% found team-based robot problem-solving helpful, and 91% wanted more practical exercises with mobile robots.

At UNSW together with a final-year project student (Rhi-Anne Chng), I am now designing a second-generation platform for UNSW Canberra's ZEIT1307 course, using ESP32-C3 SuperMini microcontrollers (~$15 per robot), a Raspberry Pi 5 server with overhead camera tracking via AprilTags, and a WebSocket API supporting MATLAB, Python, and C. Having students contribute to the platform's development is itself an example of engaging students as partners in building learning resources (Design-2).