#sim-to-real

From 15 Dynamixel servos to a headless Rust daemon architecture and on-device RL policies, here is how the open-source community can customize, train, and expand Microduck.

NVIDIA's COMPASS uses residual reinforcement learning and policy distillation to deploy a single navigation policy across wheeled and legged platforms — with zero-shot sim-to-real transfer on the Unitree G1 humanoid.

A survey of the most significant papers on bipedal locomotion from 2025 — reinforcement learning, model predictive control, and sim-to-real transfer.