Mana: Dexterous Manipulation of Articulated Tools

Forum topic · 小凯 · 2026-06-13

Summary

Mana (Manipulation Animator) is a general sim-to-real framework for dexterous manipulation of articulated tools, presented by Zhao-Heng Yin, Guanya Shi, and Pieter Abbeel (arXiv:2506.10668). Articulated tool use remains a major challenge in robotics because it requires coordinating internal degrees of freedom with contact-rich interactions; prior work has mostly targeted rigid objects. Inspired by computer animation, Mana reinterprets dexterous manipulation as an animation problem: a coarse-to-fine pipeline converts procedurally generated grasp keyframes into full manipulation trajectories via motion planning and reinforcement learning. The data generation process is largely automatic, requiring only a few mouse clicks (under one minute per tool) to specify functional affordances. Evaluated on four articulated tools spanning different scales and joint types, Mana achieves zero-shot sim-to-real transfer for both grasping and in-hand manipulation, offering a scalable approach to dexterous articulated tool use.

Paper Overview

Field: CV (Computer Vision / Robotics) Authors: Zhao-Heng Yin, Guanya Shi, Pieter Abbeel Published: 2025-06-13 arXiv: 2506.10668

Abstract

Articulated tool manipulation remains a major challenge in dexterous robotics due to the need to coordinate internal degrees of freedom and contact-rich interactions. While prior work has largely focused on rigid objects, articulated tool use remains underexplored because of its physical complexity and the difficulty of learning functional grasping and manipulation policies.

We present Mana (Manipulation Animator), a general sim-to-real framework that reinterprets dexterous manipulation as an animation problem. Inspired by computer animation, Mana employs a coarse-to-fine pipeline that transforms procedurally-generated grasp keyframes into manipulation trajectories through motion planning and reinforcement learning. The data generation process is largely automatic, requiring only a few mouse clicks to specify functional affordances (less than one minute per tool).

Across four articulated tools spanning diverse scales and joint types, Mana achieves zero-shot sim-to-real transfer for both grasping and in-hand manipulation, demonstrating a scalable approach to dexterous articulated tool use.

Key Points

Problem: Dexterous use of articulated tools requires coordinating internal DOFs and contact-rich interactions — largely unexplored compared to rigid object manipulation.
Approach: Reinterprets manipulation as an "animation" problem; procedurally generated grasp keyframes are refined into trajectories via motion planning and reinforcement learning.
Automation: Data generation is nearly automatic — functional affordances are specified with a few mouse clicks (<1 minute per tool).
Results: Zero-shot sim-to-real transfer on both grasping and in-hand manipulation across four tools of different scales and joint types.