FieldGen: From Teleoperated Pre-Manipulation Trajectories to Field-Guided Data Generation
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866912673503903744 |
|---|---|
| author | Wang, Wenhao Ye, Kehe Zhou, Xinyu Chen, Tianxing Min, Cao Zhu, Qiaoming Yang, Xiaokang Luo, Ping Shen, Yongjian Yang, Yang Yao, Maoqing Mu, Yao |
| author_facet | Wang, Wenhao Ye, Kehe Zhou, Xinyu Chen, Tianxing Min, Cao Zhu, Qiaoming Yang, Xiaokang Luo, Ping Shen, Yongjian Yang, Yang Yao, Maoqing Mu, Yao |
| contents | Large-scale and diverse datasets are vital for training robust robotic manipulation policies, yet existing data collection methods struggle to balance scale, diversity, and quality. Simulation offers scalability but suffers from sim-to-real gaps, while teleoperation yields high-quality demonstrations with limited diversity and high labor cost. We introduce FieldGen, a field-guided data generation framework that enables scalable, diverse, and high-quality real-world data collection with minimal human supervision. FieldGen decomposes manipulation into two stages: a pre-manipulation phase, allowing trajectory diversity, and a fine manipulation phase requiring expert precision. Human demonstrations capture key contact and pose information, after which an attraction field automatically generates diverse trajectories converging to successful configurations. This decoupled design combines scalable trajectory diversity with precise supervision. Moreover, FieldGen-Reward augments generated data with reward annotations to further enhance policy learning. Experiments demonstrate that policies trained with FieldGen achieve higher success rates and improved stability compared to teleoperation-based baselines, while significantly reducing human effort in long-term real-world data collection. Webpage is available at https://fieldgen.github.io/. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2510_20774 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | FieldGen: From Teleoperated Pre-Manipulation Trajectories to Field-Guided Data Generation Wang, Wenhao Ye, Kehe Zhou, Xinyu Chen, Tianxing Min, Cao Zhu, Qiaoming Yang, Xiaokang Luo, Ping Shen, Yongjian Yang, Yang Yao, Maoqing Mu, Yao Robotics Artificial Intelligence Human-Computer Interaction Large-scale and diverse datasets are vital for training robust robotic manipulation policies, yet existing data collection methods struggle to balance scale, diversity, and quality. Simulation offers scalability but suffers from sim-to-real gaps, while teleoperation yields high-quality demonstrations with limited diversity and high labor cost. We introduce FieldGen, a field-guided data generation framework that enables scalable, diverse, and high-quality real-world data collection with minimal human supervision. FieldGen decomposes manipulation into two stages: a pre-manipulation phase, allowing trajectory diversity, and a fine manipulation phase requiring expert precision. Human demonstrations capture key contact and pose information, after which an attraction field automatically generates diverse trajectories converging to successful configurations. This decoupled design combines scalable trajectory diversity with precise supervision. Moreover, FieldGen-Reward augments generated data with reward annotations to further enhance policy learning. Experiments demonstrate that policies trained with FieldGen achieve higher success rates and improved stability compared to teleoperation-based baselines, while significantly reducing human effort in long-term real-world data collection. Webpage is available at https://fieldgen.github.io/. |
| title | FieldGen: From Teleoperated Pre-Manipulation Trajectories to Field-Guided Data Generation |
| topic | Robotics Artificial Intelligence Human-Computer Interaction |
| url | https://arxiv.org/abs/2510.20774 |