RL training environments for Microduck

Pollen Robotics has published microduck_rl on GitHub, a set of reinforcement learning training environments for Microduck, a bipedal robot about 25 cm tall weighing roughly 800 g. The environments are built on mjlab (MuJoCo Warp) and use PPO. Policies train at 50 Hz, export to ONNX, and run on the real robot through the separate […]