On the Feasibility of Learning Finger-gaiting In-hand Manipulation with Intrinsic Sensing
This work addresses the problem of enabling robots to perform complex in-hand manipulation tasks with limited sensing, which is incremental as it builds on existing RL methods for dexterous manipulation.
The paper tackles the challenge of learning finger-gaiting in-hand manipulation with intrinsic sensing, using model-free reinforcement learning to achieve large-angle object re-orientation via precision grasps, resulting in significantly improved sample complexity and robust policies that transfer to novel objects.
Finger-gaiting manipulation is an important skill to achieve large-angle in-hand re-orientation of objects. However, achieving these gaits with arbitrary orientations of the hand is challenging due to the unstable nature of the task. In this work, we use model-free reinforcement learning (RL) to learn finger-gaiting only via precision grasps and demonstrate finger-gaiting for rotation about an axis purely using on-board proprioceptive and tactile feedback. To tackle the inherent instability of precision grasping, we propose the use of initial state distributions that enable effective exploration of the state space. Our method can learn finger-gaiting with significantly improved sample complexity than the state-of-the-art. The policies we obtain are robust and also transfer to novel objects. Videos can be found at https://roamlab.github.io/learnfg/