KineVLA: Towards Kinematics-Aware Vision-Language-Action Models with Bi-Level Action Decomposition
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_ | 1866917351465680896 |
|---|---|
| author | Han, Gaoge Gao, Zhengqing Li, Ziwen Huang, Jiaxin Huang, Shaoli Karray, Fakhri Gong, Mingming Liu, Tongliang |
| author_facet | Han, Gaoge Gao, Zhengqing Li, Ziwen Huang, Jiaxin Huang, Shaoli Karray, Fakhri Gong, Mingming Liu, Tongliang |
| contents | In this paper, we introduce a novel kinematics-rich vision-language-action (VLA) task, in which language commands densely encode diverse kinematic attributes (such as direction, trajectory, orientation, and relative displacement) from initiation through completion, at key moments, unlike existing action instructions that capture kinematics only coarsely or partially, thereby supporting fine-grained and personalized manipulation. In this setting, where task goals remain invariant while execution trajectories must adapt to instruction-level kinematic specifications. To address this challenge, we propose KineVLA, a vision-language-action framework that explicitly decouples goal-level invariance from kinematics-level variability through a bi-level action representation and bi-level reasoning tokens to serve as explicit, supervised intermediate variables that align language and action. To support this task, we construct the kinematics-aware VLA datasets spanning both simulation and real-world robotic platforms, featuring instruction-level kinematic variations and bi-level annotations. Extensive experiments on LIBERO and a Realman-75 robot demonstrate that KineVLA consistently outperforms strong VLA baselines on kinematics-sensitive benchmarks, achieving more precise, controllable, and generalizable manipulation behaviors. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2603_17524 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | KineVLA: Towards Kinematics-Aware Vision-Language-Action Models with Bi-Level Action Decomposition Han, Gaoge Gao, Zhengqing Li, Ziwen Huang, Jiaxin Huang, Shaoli Karray, Fakhri Gong, Mingming Liu, Tongliang Robotics Artificial Intelligence In this paper, we introduce a novel kinematics-rich vision-language-action (VLA) task, in which language commands densely encode diverse kinematic attributes (such as direction, trajectory, orientation, and relative displacement) from initiation through completion, at key moments, unlike existing action instructions that capture kinematics only coarsely or partially, thereby supporting fine-grained and personalized manipulation. In this setting, where task goals remain invariant while execution trajectories must adapt to instruction-level kinematic specifications. To address this challenge, we propose KineVLA, a vision-language-action framework that explicitly decouples goal-level invariance from kinematics-level variability through a bi-level action representation and bi-level reasoning tokens to serve as explicit, supervised intermediate variables that align language and action. To support this task, we construct the kinematics-aware VLA datasets spanning both simulation and real-world robotic platforms, featuring instruction-level kinematic variations and bi-level annotations. Extensive experiments on LIBERO and a Realman-75 robot demonstrate that KineVLA consistently outperforms strong VLA baselines on kinematics-sensitive benchmarks, achieving more precise, controllable, and generalizable manipulation behaviors. |
| title | KineVLA: Towards Kinematics-Aware Vision-Language-Action Models with Bi-Level Action Decomposition |
| topic | Robotics Artificial Intelligence |
| url | https://arxiv.org/abs/2603.17524 |