AnchorDP3: 3D Affordance Guided Sparse Diffusion Policy for Robotic Manipulation

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Zhao, Ziyan, Fan, Ke, Xu, He-Yang, Qiao, Ning, Peng, Bo, Gao, Wenlong, Li, Dongjiang, Shen, Hui
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866918069637480448
author Zhao, Ziyan
Fan, Ke
Xu, He-Yang
Qiao, Ning
Peng, Bo
Gao, Wenlong
Li, Dongjiang
Shen, Hui
author_facet Zhao, Ziyan
Fan, Ke
Xu, He-Yang
Qiao, Ning
Peng, Bo
Gao, Wenlong
Li, Dongjiang
Shen, Hui
contents We present AnchorDP3, a diffusion policy framework for dual-arm robotic manipulation that achieves state-of-the-art performance in highly randomized environments. AnchorDP3 integrates three key innovations: (1) Simulator-Supervised Semantic Segmentation, using rendered ground truth to explicitly segment task-critical objects within the point cloud, which provides strong affordance priors; (2) Task-Conditioned Feature Encoders, lightweight modules processing augmented point clouds per task, enabling efficient multi-task learning through a shared diffusion-based action expert; (3) Affordance-Anchored Keypose Diffusion with Full State Supervision, replacing dense trajectory prediction with sparse, geometrically meaningful action anchors, i.e., keyposes such as pre-grasp pose, grasp pose directly anchored to affordances, drastically simplifying the prediction space; the action expert is forced to predict both robot joint angles and end-effector poses simultaneously, which exploits geometric consistency to accelerate convergence and boost accuracy. Trained on large-scale, procedurally generated simulation data, AnchorDP3 achieves a 98.7% average success rate in the RoboTwin benchmark across diverse tasks under extreme randomization of objects, clutter, table height, lighting, and backgrounds. This framework, when integrated with the RoboTwin real-to-sim pipeline, has the potential to enable fully autonomous generation of deployable visuomotor policies from only scene and instruction, totally eliminating human demonstrations from learning manipulation skills.
format Preprint
id arxiv_https___arxiv_org_abs_2506_19269
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle AnchorDP3: 3D Affordance Guided Sparse Diffusion Policy for Robotic Manipulation
Zhao, Ziyan
Fan, Ke
Xu, He-Yang
Qiao, Ning
Peng, Bo
Gao, Wenlong
Li, Dongjiang
Shen, Hui
Robotics
Artificial Intelligence
We present AnchorDP3, a diffusion policy framework for dual-arm robotic manipulation that achieves state-of-the-art performance in highly randomized environments. AnchorDP3 integrates three key innovations: (1) Simulator-Supervised Semantic Segmentation, using rendered ground truth to explicitly segment task-critical objects within the point cloud, which provides strong affordance priors; (2) Task-Conditioned Feature Encoders, lightweight modules processing augmented point clouds per task, enabling efficient multi-task learning through a shared diffusion-based action expert; (3) Affordance-Anchored Keypose Diffusion with Full State Supervision, replacing dense trajectory prediction with sparse, geometrically meaningful action anchors, i.e., keyposes such as pre-grasp pose, grasp pose directly anchored to affordances, drastically simplifying the prediction space; the action expert is forced to predict both robot joint angles and end-effector poses simultaneously, which exploits geometric consistency to accelerate convergence and boost accuracy. Trained on large-scale, procedurally generated simulation data, AnchorDP3 achieves a 98.7% average success rate in the RoboTwin benchmark across diverse tasks under extreme randomization of objects, clutter, table height, lighting, and backgrounds. This framework, when integrated with the RoboTwin real-to-sim pipeline, has the potential to enable fully autonomous generation of deployable visuomotor policies from only scene and instruction, totally eliminating human demonstrations from learning manipulation skills.
title AnchorDP3: 3D Affordance Guided Sparse Diffusion Policy for Robotic Manipulation
topic Robotics
Artificial Intelligence
url https://arxiv.org/abs/2506.19269