BiPreManip: Learning Affordance-Based Bimanual Preparatory Manipulation through Anticipatory Collaboration

Yan Shen, Feng Jiang, Zichen He, Xiaoqi Li, Yuchen Liu, Zhiyu Li, Ruihai Wu, Hao Dong · 2026

Read BiPreManip: Learning Affordance-Based Bimanual Preparatory Manipulation through Anticipatory Collaboration by Yan Shen, Feng Jiang, Zichen He, Xiaoqi Li, Yuchen Liu, Zhiyu Li, Ruihai Wu, Hao Dong (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2603-21679.txt.

Description

Shen and co-authors teach two robot arms to cooperate on objects that one arm cannot handle directly. A flat tablet must first be pushed to a table edge; a pen cap needs the body lifted first. Their framework starts from the final goal action and works backward. A visual affordance model envisions that end state, then guides one arm through preparatory moves that set up its partner. This anticipatory planning spans long horizons and varied object types. In simulation and real trials, success rates and generalization beat strong baselines.

Catalog record

CREATORYan Shen, Feng Jiang, Zichen He, Xiaoqi Li, Yuchen Liu, Zhiyu Li, Ruihai Wu, Hao Dong
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE48970
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2603.21679

Browse the full library · About the Tower of Babel