arXiv：cs.AI· Gaoge Han, Zhengqing Gao, Ziwen Li, Jiaxin Huang, Shaoli Huang, Fakhri Karray, Mingming Gong, Tongliang Liu·· 3 小时前
ExecVLA：用双层动作表示让 VLA 模型遵循细粒度执行约束
ExecVLA: Following Fine-Grained Execution Constraints in Vision-Language-Action Models with Bi-Level Action Representation
AI 导读
ExecVLA 框架通过双层动作表示和监督式双层推理 token，将目标导向组件与执行特定组件分离，并引入目标不变性和执行可预测性目标。研究在仿真与 Realman-75 机器人上构建了执行约束遵循数据集，并在 LIBERO 和真实机器人上验证，目标完成度提升，且对指令指定执行约束的遵循更可靠。
正文
Abstract:We study fine-grained execution-constraint following in vision-language-action (VLA) models. Given an invariant task goal, the policy must follow instruction-specified execution constraints, including interaction targets, motion patterns, spatial relations, and terminal configurations. This setting exposes a limitation of goal-oriented VLAs: trajectories that complete the same task are not interchangeable when the instruction specifies how the task must be executed. We propose ExecVLA, a framework that separates a goal-oriented component from an execution-specific component through a bi-level action representation and supervised bi-level reasoning tokens. We further introduce explicit goal-invariance and execution-predictability objectives so that the goal-level representation remains stable across executions of the same goal, while the execution-level representation retains the constraints that distinguish those executions. We construct execution-constraint-following datasets in simulation and on a Realman-75 robot, with goal and fine-grained reasoning annotations. Experiments on LIBERO and the real robot show improved goal completion and, more importantly, substantially more reliable adherence to instruction-specified execution constraints.
| Subjects: | Robotics (cs.RO); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2603.17524 [cs.RO] |
| (or arXiv:2603.17524v2 [cs.RO] for this version) | |
| https://doi.org/10.48550/arXiv.2603.17524 arXiv-issued DOI via DataCite |
Submission history
From: Gaoge Han [view email]
[v1]
Wed, 18 Mar 2026 09:28:49 UTC (10,434 KB)
[v2]
Thu, 8 Oct 2026 11:17:01 UTC (10,434 KB)
来源：arXiv：cs.AI · arxiv.org