DexTransfer: Real World Multi-fingered Dexterous Grasping with Minimal Human Demonstrations
This addresses the challenge of high-dimensional dexterous grasping for robotics, offering a practical solution with reduced human input, though it appears incremental as it builds on existing dataset generation and policy learning methods.
The authors tackled the problem of teaching a multi-fingered dexterous robot to grasp objects in the real world by developing a system that uses minimal human demonstrations to generate robust grasping trajectories, achieving generalization to unseen object poses in both simulation and real-world settings.
Teaching a multi-fingered dexterous robot to grasp objects in the real world has been a challenging problem due to its high dimensional state and action space. We propose a robot-learning system that can take a small number of human demonstrations and learn to grasp unseen object poses given partially occluded observations. Our system leverages a small motion capture dataset and generates a large dataset with diverse and successful trajectories for a multi-fingered robot gripper. By adding domain randomization, we show that our dataset provides robust grasping trajectories that can be transferred to a policy learner. We train a dexterous grasping policy that takes the point clouds of the object as input and predicts continuous actions to grasp objects from different initial robot states. We evaluate the effectiveness of our system on a 22-DoF floating Allegro Hand in simulation and a 23-DoF Allegro robot hand with a KUKA arm in real world. The policy learned from our dataset can generalize well on unseen object poses in both simulation and the real world