从人类演示中学习灵巧手的连续抓取函数
Learning Continuous Grasping Function with a Dexterous Hand from Human Demonstrations
AI总结：
本文提出 CGF，用条件变分自编码器从经重定向的 3D 人类演示中学习连续时间灵巧手抓取运动，并通过仿真筛选迁移至真实 Allegro Hand，提升了多物体泛化与抓取成功率。
摘要：
我们提出利用隐式函数学习生成灵巧手操作的抓取运动。借助连续时间输入，该模型能够生成连续且平滑的抓取规划。我们将所提出的模型命名为 Continuous Grasping Function（CGF，连续抓取函数）。CGF 通过使用 Conditional Variational Autoencoder（条件变分自编码器）的生成建模方法，并利用 3D 人类演示进行学习。我们首先通过运动重定向将大规模人-物交互轨迹转换为机器人演示，随后使用这些演示训练 CGF。在推理阶段，我们在仿真器中使用 CGF 进行采样，以生成不同的抓取规划，并选择成功的规划迁移到真实机器人。通过在多样化的人类数据上训练，我们的 CGF 能够泛化以操作多个物体。与以往的规划算法相比，CGF 更加高效，并且在迁移到真实 Allegro Hand 进行抓取时，在成功率上取得了显著提升。我们的项目页面位于 https://jianglongye.com/cgf 。
英文摘要：
We propose to learn to generate grasping motion for manipulation with a dexterous hand using implicit functions. With continuous time inputs, the model can generate a continuous and smooth grasping plan. We name the proposed model Continuous Grasping Function (CGF). CGF is learned via generative modeling with a Conditional Variational Autoencoder using 3D human demonstrations. We will first convert the large-scale human-object interaction trajectories to robot demonstrations via motion retargeting, and then use these demonstrations to train CGF. During inference, we perform sampling with CGF to generate different grasping plans in the simulator and select the successful ones to transfer to the real robot. By training on diverse human data, our CGF allows generalization to manipulate multiple objects. Compared to previous planning algorithms, CGF is more efficient and achieves significant improvement on success rate when transferred to grasping with the real Allegro Hand. Our project page is available at https://jianglongye.com/cgf .
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。
