arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2602.09767cs.RO
收藏

通过无监督学习实现四足机器人的多样化技能发现

Diverse Skill Discovery for Quadruped Robots via Unsupervised Learning

College of Intelligent Robotics and Advanced Manufacturing(智能机器人与先进制造学院);Fudan University(复旦大学);College of Foreign Language(外国语言学院);The University of Shanghai for Science and Technology(上海科技大学)

Ruopeng Cui, Yifei Bi, Haojie Luo, Wei Li

AI总结:

本研究提出OMoE架构和多判别器框架,通过无监督学习实现四足机器人多样化技能发现,提升训练效率并扩大状态空间覆盖范围。

摘要:

强化学习需要专家精心设计奖励塑造以激发目标行为,而模仿学习依赖于昂贵的特定任务数据。相比之下,无监督技能发现可以通过内在动机学习多样化的有用技能,从而减少这些负担。然而,现有方法存在两个关键限制:它们通常依赖单一策略掌握多样化的行为 repertoire,而不建模其中的共享结构或区别,导致学习效率低下;此外,它们容易受到奖励黑客攻击,其中奖励信号迅速增加和收敛,而学习的技能实际多样性不足。在本工作中,我们引入了正交混合专家(OMoE)架构,以防止多样化行为陷入重叠表示,使单一策略能够掌握广泛的运动技能。此外,我们设计了多判别器框架,其中不同的判别器操作于不同的观察空间,有效缓解了奖励黑客问题。我们在12自由度Unitree A1四足机器人上评估了我们的方法,展示了多样化的运动技能集。我们的实验表明,所提出的框架提高了训练效率,并与基线相比使状态空间覆盖范围扩大了18.3%。

英文摘要:

Reinforcement learning necessitates meticulous reward shaping by specialists to elicit target behaviors, while imitation learning relies on costly task-specific data. In contrast, unsupervised skill discovery can potentially reduce these burdens by learning a diverse repertoire of useful skills driven by intrinsic motivation. However, existing methods exhibit two key limitations: they typically rely on a single policy to master a versatile repertoire of behaviors without modeling the shared structure or distinctions among them, which results in low learning efficiency; moreover, they are susceptible to reward hacking, where the reward signal increases and converges rapidly while the learned skills display insufficient actual diversity. In this work, we introduce an Orthogonal Mixture-of-Experts (OMoE) architecture that prevents diverse behaviors from collapsing into overlapping representations, enabling a single policy to master a wide spectrum of locomotion skills. In addition, we design a multi-discriminator framework in which different discriminators operate on distinct observation spaces, effectively mitigating reward hacking. We evaluated our method on the 12-DOF Unitree A1 quadruped robot, demonstrating a diverse set of locomotion skills. Our experiments demonstrate that the proposed framework boosts training efficiency and yields an 18.3\% expansion in state-space coverage compared to the baseline.

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。

↑