TESTUDO: A Robotics Kit Evolving into an AI-Driven Companion
Kaan Tabag, Mehmet Bener, Emre Dayangac, Pelin Başyurt, Irmak Üreten, Sedat Yalçın · Constructionism Conference Proceedings · 2025
Despite Artificial Intelligence (AI)-driven STEM kits being widely available, K-12 students often struggle to access them at reasonable prices. Given the crucial role of hands-on projects in early engineering education, we developed TESTUDO, an affordable, open-source, AI-integrated robotic companion[1]. It is inspired by a tortoise in design and powered by an AI-integrated low-cost microcontroller. It offers a customizable learning experience, making adaptive educational kits accessible to students from all backgrounds. After students build their device, it uses a custom-trained AI model to learn from user interactions, adapting its behavior over time to create a more engaging and personalized experience. Students can then modify and personalize their version of TESTUDO to fit their needs while continuously improving in the iteration process. This modification system relies on open-source programs, designs, and electronic frameworks. The AI model adjusts the tortoise’s development path to align with individual strengths by analyzing user data collected from verbal interactions. This adaptive system enables a dynamic progression, guiding users through tasks based on their performance and engagement levels.