Beyond World-Frame Action Heads: Motion-Centric Action Frames for Vision-Language-Action Models
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2026
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866916004360093696 |
|---|---|
| author | Yang, Huoren Zhao, Jianchao Yusong, Hu Ou, Qiguan Gao, Yuyang Ke, Wei He, Yuhang Dong, SongLin Ma, Zhiheng Gong, Yihong |
| author_facet | Yang, Huoren Zhao, Jianchao Yusong, Hu Ou, Qiguan Gao, Yuyang Ke, Wei He, Yuhang Dong, SongLin Ma, Zhiheng Gong, Yihong |
| contents | Vision-Language-Action (VLA) models have advanced rapidly with stronger backbones, broader pre-training, and larger demonstration datasets, yet their action heads remain largely homogeneous: most directly predict action commands in a fixed world coordinate frame. We propose \textbf{MCF-Proto}, a lightweight action head that equips VLA policies with a Motion-Centric Action Frame (MCF) and a prototype-based action parameterization. At each step, the policy predicts a rotation $R_t \in SO(3)$, composes actions in the transformed local frame from a set of prototypes, and maps them back to the world frame for end-to-end training, using only standard demonstrations without auxiliary supervision. This simple design induces stable emergent structure. Without explicit directional labels, the learned local frames develop a stable geometric structure whose axes are strongly compatible with demonstrated end-effector motion. Meanwhile, actions in the learned representation become substantially more compact, with variation captured by fewer dominant directions and more regularly organized by shared prototypes. These structural properties translate into improved robustness, especially under geometric perturbations. Our results suggest that adding lightweight geometric and compositional structure to the action head can materially improve how VLA policies organize and generalize robotic manipulation behavior. An anonymized code repository is provided in the supplementary material. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2605_11809 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Beyond World-Frame Action Heads: Motion-Centric Action Frames for Vision-Language-Action Models Yang, Huoren Zhao, Jianchao Yusong, Hu Ou, Qiguan Gao, Yuyang Ke, Wei He, Yuhang Dong, SongLin Ma, Zhiheng Gong, Yihong Artificial Intelligence Vision-Language-Action (VLA) models have advanced rapidly with stronger backbones, broader pre-training, and larger demonstration datasets, yet their action heads remain largely homogeneous: most directly predict action commands in a fixed world coordinate frame. We propose \textbf{MCF-Proto}, a lightweight action head that equips VLA policies with a Motion-Centric Action Frame (MCF) and a prototype-based action parameterization. At each step, the policy predicts a rotation $R_t \in SO(3)$, composes actions in the transformed local frame from a set of prototypes, and maps them back to the world frame for end-to-end training, using only standard demonstrations without auxiliary supervision. This simple design induces stable emergent structure. Without explicit directional labels, the learned local frames develop a stable geometric structure whose axes are strongly compatible with demonstrated end-effector motion. Meanwhile, actions in the learned representation become substantially more compact, with variation captured by fewer dominant directions and more regularly organized by shared prototypes. These structural properties translate into improved robustness, especially under geometric perturbations. Our results suggest that adding lightweight geometric and compositional structure to the action head can materially improve how VLA policies organize and generalize robotic manipulation behavior. An anonymized code repository is provided in the supplementary material. |
| title | Beyond World-Frame Action Heads: Motion-Centric Action Frames for Vision-Language-Action Models |
| topic | Artificial Intelligence |
| url | https://arxiv.org/abs/2605.11809 |