notesum.ai

Published at December 3

Planning-Guided Diffusion Policy Learning for Generalizable Contact-Rich Bimanual Manipulation

cs.RO

cs.CV

cs.LG

Released Date: December 3, 2024

Authors: Xuanlin Li¹, Tong Zhao¹, Xinghao Zhu¹, Jiuguang Wang¹, Tao Pang¹, Kuan Fang¹

Aff.: ¹Boston Dynamics AI Institute

Arxiv: http://arxiv.org/pdf/2412.02676v1

Refer to caption

Task (In-Dist Eval)	Planner (Sim)	Policy (Sim)	Policy (Real)
Fixed 45^∘ Rotation	0.337	0.740	0.800
Random Rotation (Easy)	0.227	0.610	0.600
Random Rotation (Medium)	0.141	0.410	0.360
Random Rotation (Hard)	0.099	0.180	0.200
Random Rotation (Overall)	0.156	0.400	0.387