Humanoid-Gym: Reinforcement Learning for Humanoid Robot with Zero-Shot Sim2Real Transfer
Isaac Gym–based RL framework for humanoid locomotion with a sim-to-sim check in MuJoCo before deploying zero-shot to real humanoids; a common starting point for humanoid RL.基于 Isaac Gym 的人形机器人运动强化学习框架，先在 MuJoCo 中做 sim2sim 验证，再零样本部署到真实人形机器人，是人形 RL 的常用起点。
BibTeX
@misc{gu2024humanoid,
title = {{Humanoid-Gym: Reinforcement Learning for Humanoid Robot with Zero-Shot Sim2Real Transfer}},
author = {Xinyang Gu and Yen-Jen Wang and Jianyu Chen},
year = {2024},
eprint = {2404.05695},
archivePrefix = {arXiv},
primaryClass = {cs.RO},
url = {https://arxiv.org/abs/2404.05695},
}Sim-to-Real & Simulators仿真与 Sim-to-RealLocomotion & Humanoids运动控制与人形机器人Humanoid人形humanoidrlisaac-gymsim-to-simsim-to-real