CMG-Net: An End-to-End Contact-Based Multi-Finger Dexterous Grasping Network
This work addresses the challenge of enabling robots to grasp objects more effectively in complex settings, representing an incremental improvement with specific gains in dexterous manipulation.
The paper tackles the problem of multi-finger dexterous grasping of unknown objects in cluttered environments by proposing a novel contact-based representation that reduces prediction dimensions and accelerates learning, resulting in CMG-Net, which significantly outperforms state-of-the-art methods for three-finger robotic hands and shows strong performance when trained on synthetic data for real robots.
In this paper, we propose a novel representation for grasping using contacts between multi-finger robotic hands and objects to be manipulated. This representation significantly reduces the prediction dimensions and accelerates the learning process. We present an effective end-to-end network, CMG-Net, for grasping unknown objects in a cluttered environment by efficiently predicting multi-finger grasp poses and hand configurations from a single-shot point cloud. Moreover, we create a synthetic grasp dataset that consists of five thousand cluttered scenes, 80 object categories, and 20 million annotations. We perform a comprehensive empirical study and demonstrate the effectiveness of our grasping representation and CMG-Net. Our work significantly outperforms the state-of-the-art for three-finger robotic hands. We also demonstrate that the model trained using synthetic data performs very well for real robots.