Abstract:

Robotics education frequently fails to connect theoretical knowledge with practical application, and rarely establishes pathways for knowledge transfer across different educational levels. This paper describes a cyclical learning model that turned competition experience into curriculum redesign and community outreach, demonstrating how student achievement can drive multilevel STEM education.

The model originated when a student team from Hochschule Karlsruhe placed first at the Upper Rhine Mobile Robotics Challenge 2024 in Mulhouse, France. The winning team then worked with faculty to redesign the mandatory course “Robogistics – Robots in Logistics” in the master program “Robotics and Artificial Intelligence in Production”. The redesigned course integrated the DuckieTown platform, a modular robotics ecosystem with small autonomous vehicles (Duckiebots) navigating in a miniature urban environment with roads, traffic signs, and driving scenarios. This MIT-developed open-source system enabled the team to transform the competition challenges into a ten-week project where former winners acted as near-peer instructors for twenty-five master's students in seven teams. Challenges for the teams were lane following, obstacle detection and avoidance, and sign recognition.

All seven teams competed in a simulated competition at the end of the course, evaluated on accuracy, error rates, and completion time. Instead of only traditional lectures, students engaged in hands-on problem-solving at their own pace. They needed both technical skills and soft skills like time and project management as well as communication. One team from this course decided to participate also at the Upper Rhine Mobile Robotics Challenge in 2025. They successfully won the first place and credited their course experience as the key preparation.

The 2025 winners then created a one-day workshop for eight female high school students (ages fifteen to seventeen). They developed a didactic concept with lesson plans and tasks at different difficulty levels, then ran the workshop independently. Participants learned programming basics, the DuckieTown interface, and completed increasingly complex robotics challenges. Workshop evaluation showed participants rated it 4.38 out of 5.0, with eighty-eight percent scoring it four or above, and 100% of participants expressing interest in future robotics workshops.

The 2024 winners are again supporting the Robogistics course in the upcoming summer term 2026 as teaching assistants. The 2025 winners will provide insights from their experience and share their lessons learned with the new class members.

The model received positive feedback at all three levels. Master's students appreciated learning through practice rather than lecture, where they could immediately fix mistakes and understand real-world applications. High school participants valued the student instructors' teaching style, the hands-on activities, and watching the autonomous robots work. Student instructors gained experience in communication, teaching methods, and managing group dynamics.

Competition success, when built into curriculum and outreach, can drive educational innovation at multiple levels. This model provides a template for STEM educators looking to boost engagement, build transferable skills, and create knowledge transfer from universities to secondary schools.

Keywords:

Mobile robotics education, competition-based learning, near-peer teaching, STEM, STEM outreach, project-based learning, DuckieTown.