OATH: Object-Agnostic Tracking for Dexterous Hands, Enabling Rapid Physics-based Retargeting

Key points

• Reinforcement learning has previously been used to achieve planning and control through contact dynamics; however, standard policies require object-specific training and fail to generalize across arbitrary geometries.1
• This work reframes physics-based retargeting as a residual learning problem, where we demonstrate that zero-shot generalization across objects relies on structured state representations.1
• Specifically, formulating the observation space to include a localized contact-error signal guiding each robot link to its target coordinate on the object surface enables a single, unified residual policy to successfully grasp novel, unseen object geometries.1
• Rolling out this policy in simulation on a kinematic reference trajectory yields a physics-based retargeter, OATH.1

Sources (1)

• [1]OATH: Object-Agnostic Tracking for Dexterous Hands, Enabling Rapid Physics-based Retargeting
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 9, 10:55 PM

  Retargeting human demonstrations into dynamically feasible dexterous hand trajectories requires resolving complex contact physics, yet current methods either omit physical constraints or rely on computationally expensive sampling-based optimization in simulation.

  Reinforcement learning has previously been used to achieve planning and control through contact dynamics; however, standard policies require object-specific training and fail to generalize across arbitrary geometries.

Extractive summary: sentences quoted from the sources.