The Planning Limits of Latent World Models

📄 arXiv: 2609.39235v1 📥 PDF

作者: Ali Alrasheed, Basim Azam, Naveed Akhtar

分类: cs.RO, cs.CV

发布日期: 2026-09-30


💡 一句话要点

研究潜在世界模型的规划极限以提升机器人行为预测能力

🎯 匹配领域: 支柱一:机器人控制 (Robot Control) 支柱二:RL算法与架构 (RL & Architecture) 支柱九:具身大模型 (Embodied Foundation Models)

关键词: 世界模型 动作条件预测 自监督学习 机器人规划 视觉-语言-动作策略

📋 核心要点

  1. 现有的世界模型在预测能力上存在局限,尤其是在规划复杂行为时,未能明确其有效性和失败情况。
  2. 论文提出使用冻结的自监督视觉骨干网络构建动作条件预测器,以评估不同模型在多样化任务中的表现。
  3. 实验结果表明,世界模型在规划范围内能显著提高成功率,且使用视觉-语言-动作策略时成功率提升至77%。

📝 摘要(中文)

世界模型为机器人理解物理世界的演变和通过想象规划复杂行为提供了有希望的途径。然而,现有研究主要展示了这些模型的能力,却未明确其预测在规划中的有效性及失败之处。本文使用五种冻结的自监督视觉骨干网络,研究了基于动作条件的预测器在多样化的Meta-World操作任务和真实机器人交互中的表现。研究发现,世界模型在规划时仅在目标位于其想象轨迹内或稍微超出时才能可靠地指导动作选择。对于远期目标,纯想象的成功率为23%,而结合反馈的规划方法成功率提升至30%。

🔬 方法详解

问题定义:本文旨在探讨现有世界模型在规划复杂行为时的预测有效性及其局限性,尤其是目标距离对成功率的影响。现有方法未能明确何时预测有用,何时失败。

核心思路:通过使用五种冻结的自监督视觉骨干网络,构建动作条件预测器,评估其在不同环境下的表现,以揭示世界模型的规划极限。

技术框架:整体架构包括五种视觉骨干网络的比较,基于这些骨干网络构建的预测器用于执行Meta-World操作任务和真实机器人交互,评估其在不同目标距离下的成功率。

关键创新:本研究的创新在于系统性地分析了世界模型在不同目标距离下的预测能力,揭示了规划成功率随目标距离变化的规律,尤其是当目标距离超过五步时成功率显著下降。

关键设计:使用的五种骨干网络包括V-JEPA 2、V-JEPA 2.1、VideoMAEv2、VideoPrism和DINOv2,实验中采用了五步的回滚策略,评估了不同模型在目标距离为五至二十步时的表现。

🖼️ 关键图片

fig_0
img_1
img_2

📊 实验亮点

实验结果显示,世界模型在规划范围内的成功率显著提高,尤其是结合视觉-语言-动作策略时,成功率从65%提升至77%。在目标距离为五步时,成功率高达92%,而当目标距离增加至二十步时,成功率降至41%。

🎯 应用场景

该研究的潜在应用领域包括机器人自主导航、智能制造和人机协作等。通过提升机器人在复杂环境中的规划能力,能够更好地应对动态变化的任务需求,进而推动智能机器人技术的实际应用和发展。

📄 摘要(原文)

World models offer a promising way to help robots understand how the physical world evolves and plan complex behaviours through imagination. Yet existing studies mainly demonstrate what these models can accomplish, leaving unclear when their predictions remain useful for planning and where they fail. We study this question using action-conditioned predictors built on five frozen self-supervised visual backbones: V-JEPA 2, V-JEPA 2.1, VideoMAEv2, VideoPrism, and DINOv2. We use frozen backbones to test representations intended to transfer across environments. We evaluate these models on diverse Meta-World manipulation tasks and real-robot interactions from BridgeData V2. We find that a world model guides action selection reliably only when the goal lies within, or slightly beyond, the trajectory it imagines during planning. With five-step rollouts, the length the predictor was trained on, the world model ranks actions reliably only for targets five to ten control steps ahead, whereas task goals lie 16 to 53 steps away. Neither an 81-fold larger predictor nor longer-rollout training extends this range; the encoder affects both range and closed-loop success, with V-JEPA 2.1 performing most consistently. More fundamentally, the limit persists under perfect prediction: using the real simulator, success falls from 92% to 41% as the target moves from five to twenty steps ahead of a five-step rollout. Planning therefore requires either longer imagined trajectories or closer subgoals. For distant goals, pure imagination succeeds in 23% of episodes, planning with feedback (MPC) raises success to 30%, imagining as far as the goal to 47%, and nearby expert subgoals to 76%. Used within its plannable range, a world model can also improve a vision-language-action (VLA) policy: choosing among eight actions the VLA proposes raises its success from 65% to 77% across 16 different tasks.