论文标题
通过对抗强化学习的故障感知的强大控制
Fault-Aware Robust Control via Adversarial Reinforcement Learning
论文作者
论文摘要
与人类和动物相比，机器人的适应能力有限。但是，机器人损坏在现实世界中普遍存在，尤其是对于部署在极端环境中的机器人。机器人的脆弱性极大地限制了他们的广泛应用。我们提出了一个对抗强化学习框架，该框架在操纵任务和运动任务中都大大提高了对关节损害案例的机器人鲁棒性。在其性能较差的联合损害案例下，对代理进行迭代训练。我们在三指机器人手和四倍的机器人上验证算法。我们的算法只能在模拟中进行培训，并直接部署在真正的机器人上，而无需进行任何微调。它还证明了超过任意关节损害案件的成功率。
Robots have limited adaptation ability compared to humans and animals in the case of damage. However, robot damages are prevalent in real-world applications, especially for robots deployed in extreme environments. The fragility of robots greatly limits their widespread application. We propose an adversarial reinforcement learning framework, which significantly increases robot robustness over joint damage cases in both manipulation tasks and locomotion tasks. The agent is trained iteratively under the joint damage cases where it has poor performance. We validate our algorithm on a three-fingered robot hand and a quadruped robot. Our algorithm can be trained only in simulation and directly deployed on a real robot without any fine-tuning. It also demonstrates exceeding success rates over arbitrary joint damage cases.
