Tactile-conditioned model improves robotic hand dexterity and stability
TacBPM: A Tactile-conditioned Behavior Prior Model for Dexterous Reorientation
Robotics
Summary
Dexterous robot hands must control many fingers precisely while feeling how they touch objects. The authors created TacBPM, a model that helps robots use touch and joint feedback to better move and rotate objects inside their hands. This model speeds up learning and keeps the object stable, even with new shapes or poses. They tested it on simulated and real robots, showing it works better than older methods.
What this means in practice
- •For robotics engineers: Train more stable and efficient robot hand manipulation policies that adapt to new objects using tactile feedback.
- •For automation integrators: Deploy robots that reliably grasp, reorient, and transport diverse tools and items with enhanced in-hand dexterity.
Authors
Jie Yin, Wanli Xing, Zeyuan Zhao, Xuezhou Zhu, Zhijie Deng, Kaifeng Zhang
Abstract
Dexterous in-hand manipulation requires policies that coordinate high-DoF hand joints through intermittent, contact-rich interaction. Beyond target-orientation tracking, such policies must discover finger gaits that preserve object stability while adapting to geometry, anisotropy, pose, contact, and sensing changes. We propose \method, a tactile-conditioned behavior prior model for dexterous reorientation. \method distills multi-scale sphere specialists into a latent controller and lets downstream policies reuse the fixed tactile prior through residual latent actions, reducing renewed exploration from raw joint commands. The prior conditions on tactile-proprioceptive history so latent behavior reflects the current hand-object interaction. We evaluate arbitrary-pose transfer across anisotropic objects, commanded-axis rotation, and an arm-hand Grasp-to-AnyPose task in which the robot must grasp, lift, transport, and reach goal poses for novel tool geometries and generalized placements. Extensive experiments demonstrate that the proposed method accelerates training and enables stable policies where matched raw-action PPO remains near failure, with successful sim-to-real transfer in in-hand and arm-hand tasks.