DexMachina: Functional Retargeting for Bimanual Dexterous Manipulation

We study the problem of functional retargeting: learning dexterous manipulation policies to track object states from human hand-object demonstrations. We focus on long-horizon, bimanual tasks with articulated objects, which are challenging due to large action space, spatiotemporal discontinuities, and the embodiment gap between human and robot hands. We propose DexMachina, a novel curriculum-based algorithm: the key idea is to use virtual object controllers with decaying strength: an object is first driven automatically towards its target states, such that the policy can gradually learn to take over under motion and contact guidance. We release a simulation benchmark with a diverse set of tasks and dexterous hands, and show that DexMachina significantly outperforms baseline methods. Our algorithm and benchmark enable a functional comparison for hardware designs, and we present key findings informed by quantitative and qualitative results. With the recent surge in dexterous hand development, we hope this work will provide a useful platform for identifying desirable hardware capabilities and lower the barrier for contributing to future research. Videos and more at \url{project-dexmachina.github.io}

We teach dexterous robot hands to perform human hand skills using reinforcement learning in simulation. We do so by using 3D-captured human hand and object interaction dataset that includes long trajectories of both human hand poses and object meshes, then we use these to create tasks in physics simulation and train a deep reinforcement learning policy to control the dexterous robot hands to recreate the same interactions.