CoFreeVLA: Collision-Free Dual-Arm Manipulation via Vision-Language-Action Model and Risk Estimation
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866918320384507904 |
|---|---|
| author | Zhai, Xuanran Ou, Binkai Yu, Qiaojun Hao, Ce Liu, Yaohua |
| author_facet | Zhai, Xuanran Ou, Binkai Yu, Qiaojun Hao, Ce Liu, Yaohua |
| contents | Vision Language Action (VLA) models enable instruction following manipulation, yet dualarm deployment remains unsafe due to under modeled selfcollisions between arms and grasped objects. We introduce CoFreeVLA, which augments an endtoend VLA with a short horizon selfcollision risk estimator that predicts collision likelihood from proprioception, visual embeddings, and planned actions. The estimator gates risky commands, recovers to safe states via risk-guided adjustments, and shapes policy refinement for safer rollouts. It is pre-trained with model-based collision labels and posttrained on real robot rollouts for calibration. On five bimanual tasks with the PiPER robot arm, CoFreeVLA reduces selfcollisions and improves success rates versus RDT and APEX. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2601_21712 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | CoFreeVLA: Collision-Free Dual-Arm Manipulation via Vision-Language-Action Model and Risk Estimation Zhai, Xuanran Ou, Binkai Yu, Qiaojun Hao, Ce Liu, Yaohua Robotics Vision Language Action (VLA) models enable instruction following manipulation, yet dualarm deployment remains unsafe due to under modeled selfcollisions between arms and grasped objects. We introduce CoFreeVLA, which augments an endtoend VLA with a short horizon selfcollision risk estimator that predicts collision likelihood from proprioception, visual embeddings, and planned actions. The estimator gates risky commands, recovers to safe states via risk-guided adjustments, and shapes policy refinement for safer rollouts. It is pre-trained with model-based collision labels and posttrained on real robot rollouts for calibration. On five bimanual tasks with the PiPER robot arm, CoFreeVLA reduces selfcollisions and improves success rates versus RDT and APEX. |
| title | CoFreeVLA: Collision-Free Dual-Arm Manipulation via Vision-Language-Action Model and Risk Estimation |
| topic | Robotics |
| url | https://arxiv.org/abs/2601.21712 |