Sim-to-Real of Humanoid Locomotion Policies via Joint Torque Space Perturbation Injection

This paper proposes a novel alternative to existing sim-to-real methods for training control policies with simulated experiences. Prior sim-to-real methods for legged robots mostly rely on the domain randomization approach, where a fixed finite set of simulation parameters is randomized during training. Instead, our method adds state-dependent perturbations to the input joint torque used for forward simulation during the training phase. These state-dependent perturbations are designed to simulate a broader range of reality gaps than those captured by randomizing a fixed set of simulation parameters. Experimental results show that our method enables humanoid locomotion policies that achieve greater robustness against complex reality gaps unseen in the training domain.

@article{cha2025_2504.06585,
  title={ Sim-to-Real of Humanoid Locomotion Policies via Joint Torque Space Perturbation Injection },
  author={ Woohyun Cha and Junhyeok Cha and Jaeyong Shin and Donghyeon Kim and Jaeheung Park },
  journal={arXiv preprint arXiv:2504.06585},
  year={ 2025 }
}