通过联合扭矩空间扰动注入实现人形机器人的仿真到现实
Sim-to-Real of Humanoid Locomotion Policies via Joint Torque Space Perturbation Injection
AI总结：
本文提出一种新的仿真到现实方法，通过在正向仿真中注入状态依赖的扰动，提高人形机器人运动策略的鲁棒性。
摘要：
本文提出了一种新的仿真到现实方法，用于训练具有模拟经验的控制策略。与以往依赖固定有限参数域随机化的做法不同，所提出的方法在正向仿真中注入状态依赖的扰动。这些扰动旨在模拟比标准参数随机化更广的现实差距，而无需额外训练。通过使用神经网络作为灵活的扰动生成器，该方法可以表示复杂的、状态依赖的不确定性，如非线性执行器动力学和接触刚度，这些是参数随机化无法捕捉的。实验结果表明，所提出的方法使人形机器人运动策略在仿真和现实部署中都能更好地应对复杂且未见过的现实差距。
英文摘要：
This paper proposes a novel alternative to existing sim-to-real methods for training control policies with simulated experiences. Unlike prior methods that typically rely on domain randomization over a fixed finite set of parameters, the proposed approach injects state-dependent perturbations into the input joint torque during forward simulation. These perturbations are designed to simulate a broader spectrum of reality gaps than standard parameter randomization without requiring additional training. By using neural networks as flexible perturbation generators, the proposed method can represent complex, state-dependent uncertainties, such as nonlinear actuator dynamics and contact compliance, that parametric randomization cannot capture. Experimental results demonstrate that the proposed approach enables humanoid locomotion policies to achieve superior robustness against complex, unseen reality gaps in both simulation and real-world deployment.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
