发表机构
LS Wiiri Robot Innovation Center; Technical University of Munich; Volkswagen Group（LS维里机器人创新中心; 慕尼黑工业大学; 大众集团）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
研究针对机器人运动生成中约束与数据不匹配问题，提出ConFlow框架，将约束信息纳入训练目标，用条件高斯过程解决设计规范，用不可行演示作负监督，实验表明其能降低碰撞率、提高轨迹质量，弥合训练与推理差距。
AI 中文摘要
近年来，流匹配已成为生成式建模机器人运动生成的一种重要方法。一般形式的流匹配是基于常微分方程的神经采样器，通过回归与运动样本相关的经验流场作为数据进行训练。然而，在机器人运动生成中，我们通常有一些在收集的数据中可能不存在的额外约束。当前大多数方法在可用数据上训练流，并在推理时使用引导来强制执行特定任务的约束。为了解决这种不匹配，我们提出了ConFlow，这是一个约束引导的流匹配框架，通过可微障碍或成本函数将约束信息直接纳入训练目标。为了解决诸如平滑性和边界条件等设计规范，我们建议用条件高斯过程替换流匹配训练中使用的标准高斯源分布。我们的方法还使用不可行的演示作为负监督，在不需要额外专家数据的情况下提高约束满足度。在双机器人导航任务上的实验表明，无论有无推理时引导，ConFlow都比标准流匹配基线实现了更低的碰撞率和更高的轨迹质量。这些结果验证了训练时约束整合是弥合生成运动模型中训练 - 推理差距的有效方法。
英文摘要
In recent years Flow Matching has become a prominent method for generative modeling robot motion generation. In its generic form Flow Matching is an ODE-based neural sampler that is trained by regressing empirical flow fields associated with motion samples as data. However, in robot motion generation we often have additional constraints that might not be present in the collected data. The majority of current approaches train the flow on the available data and use inference-time guidance to enforce task-specific constraints. To address this mismatch, we propose \textbf{ConFlow}, a constraint-guided flow matching framework that incorporates constraint information directly into the training objective via differentiable barrier or cost functions. To address design specifications such as smoothness and boundary conditions, we propose replacing the standard Gaussian source distribution used in flow matching training with a conditional Gaussian Process. Our approach also uses infeasible demonstrations as negative supervision, improving constraint satisfaction without requiring additional expert data. Experiments on a two-robot navigation task demonstrate that ConFlow achieves lower collision rates and higher trajectory quality than standard flow matching baselines, with or without inference-time guidance. These results validate training-time constraint integration as an effective approach to closing the training--inference gap in generative motion models.
CommentsRSS 2026 Workshop on Diffusion for Robot Learning
