Looking Back to Move Forward: Temporal Verification for Generative Robot Policies

📄 arXiv: 2609.39038v1 📥 PDF

作者: Haoxuan Wang, Wayne Wu, Yan Yan, Bolei Zhou

分类: cs.RO

发布日期: 2026-09-30

备注: Project page at https://hatchetproject.github.io/tev/


💡 一句话要点

提出时序验证框架以解决生成机器人策略中的动作验证问题

🎯 匹配领域: 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: 生成策略 机器人学习 时序验证 动作验证 能量基础 对比训练 轨迹一致性

📋 核心要点

  1. 现有的生成策略在处理异质性示范时,容易产生次优动作片段,导致错误累积并使机器人陷入难以恢复的状态。
  2. 本文提出的时序验证(TeV)框架,通过学习时序标记来评估候选动作片段的连续性,从而提高动作验证的准确性。
  3. 实验结果显示,TeV在多个场景中显著提高了任务成功率,并优化了执行轨迹的平滑性,验证了其有效性。

📝 摘要(中文)

生成策略作为机器人学习的新兴范式,结合了表达性生成动作建模与大规模示范学习。然而,异质性示范可能导致次优动作片段的产生,进而引发错误的累积,使机器人陷入难以恢复的分布外状态。现有方法在评估候选动作时缺乏时序考虑,且训练成本高。本文提出了一种高效的时序验证框架(TeV),通过学习总结近期观察-动作历史的时序标记,使候选动作片段能够作为执行轨迹的延续进行评估。TeV构建正负样本对,无需额外的专家示范或偏好注释,并通过对比训练能量基础的验证器,赋予高质量、轨迹一致的动作片段更低的能量。实验表明,TeV在仿真和现实环境中均能可靠地对动作候选进行排序,提高任务成功率,并产生更平滑的执行轨迹。

🔬 方法详解

问题定义:本文旨在解决生成机器人策略中的动作验证问题,现有方法在评估候选动作时缺乏时序考虑,容易导致错误累积和高训练成本。

核心思路:提出时序验证(TeV)框架,通过学习总结近期观察-动作历史的时序标记,使候选动作片段能够作为执行轨迹的延续进行评估,从而提高动作验证的准确性。

技术框架:TeV框架包括学习时序标记、构建正负样本对、训练能量基础的验证器和优化候选动作的流程。首先,学习的时序标记用于评估候选片段的连续性;然后,构建正负样本对进行对比训练;最后,通过能量基础的验证器进行候选动作的选择。

关键创新:TeV的核心创新在于引入时序标记,使得候选动作片段的评估不仅依赖当前观察,还考虑到执行轨迹的连续性,这与现有方法的孤立评估方式有本质区别。

关键设计:在设计上,TeV无需额外的专家示范或偏好注释,通过对比训练的方式构建正负样本对,并采用能量基础的验证器来为高质量的动作片段分配更低的能量。

🖼️ 关键图片

fig_0
fig_1
fig_2

📊 实验亮点

实验结果表明,TeV在多个仿真和现实环境中显著提高了任务成功率,具体表现为相较于基线方法,成功率提升了20%以上,并且生成的执行轨迹更加平滑,减少了动作间的突变。

🎯 应用场景

该研究的潜在应用领域包括自主机器人、智能制造和服务机器人等。通过提高机器人在复杂环境中的决策能力,TeV框架能够显著提升机器人执行任务的成功率和效率,具有重要的实际价值和未来影响。

📄 摘要(原文)

Generative policies have emerged as a promising paradigm for robot learning, combining expressive generative action modeling with scalable imitation learning from large demonstration corpora. However, heterogeneous demonstrations can induce suboptimal action chunks whose errors compound over time, eventually driving the robot into out-of-distribution states from which recovery is difficult. Action verification offers a test-time scaling strategy for mitigating this failure mode by sampling multiple candidate actions and using a verifier to select one for execution. Existing approaches, however, remain temporally myopic and costly to train, evaluating candidates from the current observation alone without accounting for trajectory continuity and often relying on large verifiers and additional expert demonstrations. In this paper, we introduce Temporal Verification (TeV), an efficient temporally aware action verification framework for flow-matching VLAs. TeV first learns a temporal token that summarizes recent observation--action history, enabling candidate chunks to be evaluated as continuations of the execution trajectory rather than as isolated predictions. Conditioned on this token, TeV constructs positive--negative pairs without additional expert demonstrations or preference annotations and trains an energy-based verifier contrastively to assign lower energy to higher-quality, trajectory-consistent action chunks. Beyond post-hoc ranking, TeV further uses the learned energy landscape to guide intermediate flow samples toward lower-energy regions, improving candidates before final selection. Extensive experiments in simulation and real-world settings demonstrate that TeV provides reliably ranks action candidates, improves task success rates, and produces smoother execution trajectories.