Ojaghi Pegah, Mir Romina, Marjaninejad Ali, Erwin Andrew, Wehner Michael, Valero-Cuevas Francisco J
Computer Science and Engineering Department, University of California Santa Cruz, Santa Cruz, CA, USA.
Alfred E. Mann Department of Biomedical Engineering, University of Southern California, Los Angeles, CA, USA.
Sci Adv. 2025 Apr 4;11(14):eadp8407. doi: 10.1126/sciadv.adp8407. Epub 2025 Apr 2.
Dexterous manipulation remains an aspirational goal for autonomous robotic systems, particularly when learning to lift and rotate objects against gravity with intermittent finger contacts. We use model-free reinforcement learning to compare the effect of curriculum (i.e., combinations of lift and rotation tasks) and haptic information (i.e., no-tactile versus 3D-force) on learning with a simulated three-finger robotic hand. In addition, a novel curriculum-based learning rate scheduler accelerates convergence. We demonstrate that the choice of curriculum biases the progression of learning for dexterous manipulation across objects with different weights, sizes, and shapes-underscoring the robustness of our learning approach. Unexpectedly, learning is achieved even in the absence of haptic information. This challenges conventional thinking about task "complexity" and the necessity of haptic information for dexterous manipulation for this task. This work invites the analogy of curriculum learning as a malleable developmental process from a pluripotent state driven by the nature of the learning experience.
灵巧操作仍然是自主机器人系统的一个理想目标，尤其是在学习通过间歇性手指接触来对抗重力举起和旋转物体时。我们使用无模型强化学习来比较课程（即举起和旋转任务的组合）和触觉信息（即无触觉与三维力）对使用模拟三指机器人手进行学习的影响。此外，一种新颖的基于课程的学习率调度器加速了收敛。我们证明，课程的选择会影响在不同重量、大小和形状的物体上进行灵巧操作的学习进程，这突出了我们学习方法的稳健性。出乎意料的是，即使没有触觉信息也能实现学习。这挑战了关于任务“复杂性”以及触觉信息对于此任务灵巧操作必要性的传统观念。这项工作引发了将课程学习类比为一个由学习经验的性质驱动的从多能状态开始的可塑发展过程。
