让视觉语言动作模型先规划再执行，提升复杂任务表现
Coarse-to-Control: Action-Token Planning for Vision-Language-Action Models
- 用离散动作令牌分步规划未来行为轨迹
- 在长序列任务上性能显著提升，多阶段任务增益最大
- 适合需要长期规划的机器人操作场景
大多数视觉-语言-动作（VLA）模型直接从观测映射到动作，缺乏显式中间规划，导致在长时程任务中早期错误累积。我们提出 Coarse-to-Control，一种原生在动作令牌空间中进行规划的计划-执行型VLA。核心思想是让策略首先预测一个紧凑的粗粒度动作令牌序列，以概括预期的未来轨迹，然后基于该计划生成可执行的动作令牌。由于规划与执行共享统一的离散动作词汇表，计划始终贴近控制流形，提供可直接使用的指导，而非需转换回运动指令的抽象提示。在LIBERO、SimplerEnv-WidowX及真实世界操作任务上的实验表明，动作令牌规划持续优于直接动作生成，尤其在长时程多阶段任务中收益最大。
原文摘要 · Abstract (English)
Most vision-language-action (VLA) models map observations directly to actions without explicit intermediate planning, which limits performance on long-horizon tasks where early mistakes compound. We propose Coarse-to-Control, a plan-execute VLA that introduces planning natively in the action-token space. The key idea is to let the policy first predict a compact sequence of coarse action tokens that summarize the intended future trajectory, and then generate executable action tokens conditioned on this plan. Because both planning and execution share a unified discrete action vocabulary, the plan stays close to the control manifold and provides directly actionable guidance rather than an abstract hint that must be translated back to motor commands. Experiments on LIBERO, SimplerEnv-WidowX, and real-world manipulation tasks show that action-token planning consistently improves over direct action generation, with the largest gains on long-horizon multi-stage tasks.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。