Duckietown Robots :: Vision-Based Autonomous Robot Platform Designed for Students, Educators & Researchers Worldwide

Duckietown is a Swiss-based robotics education and research company best known for its innovative, low-cost, vision-based autonomous robot platform designed to make robotics accessible to students, educators, and researchers worldwide. Originally developed at MIT under the leadership of Professor Daniela Rus, the Duckietown project has evolved into a full-fledged company (Duckietown AG) headquartered in Switzerland. The core philosophy is “learning by doing” through hands-on, affordable robotics that emphasizes computer vision, AI, and autonomous navigation without relying on expensive sensors like LiDAR.

The flagship product is the Duckiebot, a small, duck-themed differential-drive robot that uses a monocular camera as its primary sensor. Students assemble and program these robots to navigate city-like environments (called “Duckietowns”) made of road-like floor mats with colored tape lanes, traffic signs, and obstacles. The latest generation, Duckiebot DB21, features a Raspberry Pi 4, improved cameras, better motors, and modern software support.

Complementing the robots is the Duckietown MOOC (Massive Open Online Course) curriculum, used by hundreds of universities globally. The complete ecosystem includes:

• Duckiebot robots (DB21, Jetbot Edition, etc.)
• Duckietown city kits with roads, intersections, traffic lights, and buildings
• Simulator (Gym-Duckietown)
• Fleet management tools for running multiple robots simultaneously
• Advanced perception and planning software stack based on ROS, PyTorch, and Docker

Duckietown is used in over 300 universities across 60+ countries and has become a standard platform for teaching robot perception, control, reinforcement learning, and multi-agent systems. The company also offers professional development kits, teacher training, and customized corporate workshops. Their hardware is deliberately designed to be affordable (typically under $300 per robot), repairable, and open-source, aligning with their mission to democratize robotics education.

The company maintains a strong open-source culture, with all software, documentation, and many hardware designs available on GitHub. Duckietown’s approach has been particularly successful because it focuses on real-world computer-vision challenges (lane following, object detection, semantic segmentation) rather than abstract theoretical problems, giving students immediately tangible and exciting results.