ManipTrans: Efficient Dexterous Bimanual Manipulation Transfer via Residual Learning
This work addresses the problem of data scarcity for dexterous robotic manipulation, enabling more efficient skill transfer and dataset creation for researchers in robotics and embodied AI.
The paper tackles the challenge of transferring human bimanual skills to dexterous robotic hands by introducing ManipTrans, a two-stage method that pre-trains a trajectory imitator and fine-tunes a residual module, resulting in improved success rates, fidelity, and efficiency over state-of-the-art methods and enabling the creation of a large-scale dataset DexManipNet with 3.3K episodes.
Human hands play a central role in interacting, motivating increasing research in dexterous robotic manipulation. Data-driven embodied AI algorithms demand precise, large-scale, human-like manipulation sequences, which are challenging to obtain with conventional reinforcement learning or real-world teleoperation. To address this, we introduce ManipTrans, a novel two-stage method for efficiently transferring human bimanual skills to dexterous robotic hands in simulation. ManipTrans first pre-trains a generalist trajectory imitator to mimic hand motion, then fine-tunes a specific residual module under interaction constraints, enabling efficient learning and accurate execution of complex bimanual tasks. Experiments show that ManipTrans surpasses state-of-the-art methods in success rate, fidelity, and efficiency. Leveraging ManipTrans, we transfer multiple hand-object datasets to robotic hands, creating DexManipNet, a large-scale dataset featuring previously unexplored tasks like pen capping and bottle unscrewing. DexManipNet comprises 3.3K episodes of robotic manipulation and is easily extensible, facilitating further policy training for dexterous hands and enabling real-world deployments.