论文/arXiv 2605.19580·5 个月前

PAPO-VLA：规划感知策略优化提升VLA模型操作可靠性

针对VLA模型闭环交互不可靠难题，提出PAPO-VLA算法。通过因果充分性与必要性量化规划动作重要性并融入GRPO优势估计，在多基准测试中验证了其对任务成功率的有效提升。

AI 整理

• 具身模型

本文由 AI 根据公开内容整理，可能有误，请以原文为准：PAPO-VLA: Planning-Aware Policy Optimization for Vision-Language-Action Models（arXiv 2605.19580）

核心内容

针对视觉-语言-动作（VLA）模型在语言引导机器人操作中因闭环交互导致可靠性不足的问题，arXiv论文提出规划感知策略优化方法PAPO-VLA。该研究指出VLA策略兼具改变执行方向的“规划器”与实现密集连续动作的“执行器”双重角色，现有模仿学习或轨迹优化方法缺乏对关键规划动作的显式识别与重要性度量，导致策略在长程任务中易失效。

PAPO-VLA技术方案首先联合动作变化率与轨迹结果识别规划动作，进而利用因果充分性与因果必要性估算其重要性权重，并将该权重注入GRPO优势估计函数中。该机制使高重要性规划动作获得更强优化梯度，同时保留轨迹级反馈以维持整体执行质量。实验表明，该方法在多个机器人操作基准上有效提升了VLA模型的决策准确性与任务完成可靠性。

关键看点

• 双重角色解耦：将VLA策略分解为任务导向规划器与密集动作执行器，明确规划动作对后续执行路径的决定性影响
• 因果重要性度量：基于因果充分性与必要性联合评估规划动作权重，解决传统方法无法量化单步决策对任务成功贡献度的问题
• GRPO优势增强：将规划动作重要性融入GRPO优势估计，实现关键决策点的差异化梯度更新，兼顾局部规划与全局轨迹优化

工程与产业观察

PAPO-VLA为具身大模型训练提供了从“端到端黑盒”向“结构化归因”演进的新范式。通过引入因果推断量化规划价值，解决了VLA在复杂操作中长程依赖与信用分配难题，对提升机器人在非结构化环境中的自主决策鲁棒性具有直接工程参考价值。

英文摘要 (Abstract)

Vision-Language-Action (VLA) models show promising ability in language-guided robotic tasks. However, making VLA policies reliable remains challenging, because a manipulation task is completed through closed-loop interaction, where each action affects subsequent execution. To analyze this problem, we revisit VLA policy during execution and argue that a VLA policy acts both as a planner, which makes task-oriented decisions that change the direction of execution, and as an executor, which realizes these decisions through dense continuous actions. This view suggests that improving VLA reliability requires particular attention to planning actions. Existing optimization methods can imitate actions or improve complete trajectories, but they usually do not explicitly identify planning actions or measure their importance for task success. To address this issue, we propose Planning-Aware Policy Optimization for VLA models (PAPO-VLA). PAPO-VLA first identifies planning actions by jointly considering action variation and trajectory outcome, then estimates their importance through causal sufficiency and causal necessity, and finally incorporates this importance into GRPO advantage estimation. In this way, more important planning actions receive stronger optimization emphasis, while the whole trajectory is still optimized by trajectory-level feedback. Experiments on multiple benchmarks demonstrate the effectiveness of PAPO-VLA.

作者：Peizheng Guo, Jingyao Wang, Changwen Zheng, Wenwen Qiang

来源

本页事实对应的公开页面，以原文为准。

1. arXiv 2605.19580：PAPO-VLA: Planning-Aware Policy Optimization for Vision-Language-Action Models访问 2026-10-23