Balancing and push-recovery are essential capabilities enabling humanoid robots to solve complex locomotion tasks. In this context, classical control systems tend to be based on simplified physical models and hard-coded strategies. Although successful in specific scenarios, this approach requires demanding tuning of parameters and switching logic between specifically-designed controllers for handling more general perturbations. We apply model-free Deep Reinforcement Learning for training a general and robust humanoid push-recovery policy in a simulation environment. Our method targets high-dimensional whole-body humanoid control and is validated on the iCub humanoid. Reward components incorporating expert knowledge on humanoid control enable fast learning of several robust behaviors by the same policy, spanning the entire body. We validate our method with extensive quantitative analyses in simulation, including out-of-sample tasks which demonstrate policy robustness and generalization, both key requirements towards real-world robot deployment.

中文翻译：

---

从类人机器人推挽式恢复学习谈整体策略的出现

平衡和推回恢复是必不可少的功能，可让类人机器人来解决复杂的运动任务。在这种情况下，经典的控制系统往往基于简化的物理模型和硬编码策略。尽管在特定情况下是成功的，但是这种方法要求对参数进行微调，并且需要在专门设计的控制器之间切换逻辑以处理更一般的扰动。我们将无模型的深度强化学习应用于在模拟环境中训练通用且鲁棒的类人动物推挽式恢复策略。我们的方法针对高维全身类人动物控制，并在iCub类人动物上得到了验证。奖励组件结合了类人动物控制方面的专业知识，可通过同一策略快速学习多个健壮的行为，遍及整个身体。