Microduck puts reinforcement learning into a 25-centimetre robot
Pollen Robotics’ small open-source biped is designed as an affordable platform for learning, experimentation and embodied AI.
The story
Pollen Robotics has opened pre-orders for Microduck, a 25-centimetre open-source bipedal robot priced from $399.
The platform uses reinforcement-learning techniques and is intended to give students, makers and researchers a practical way to explore locomotion and robot control without a full-size laboratory system.
Its significance is accessibility. Lower-cost hardware can widen participation in embodied AI, while an open design lets the community inspect, modify and share improvements rather than treating the machine as a closed consumer product.
INNOVOX analysis
Affordable open hardware can create the robotics equivalent of an educational computer: capable enough for real experimentation and transparent enough to modify. Community documentation will be crucial to its impact.
What to watch
Shipping quality, replacement parts, curriculum support and the range of reproducible projects will determine whether Microduck becomes a lasting platform.