静态输入，动态输出：面向运动物体操作的反事实动作增强方法
Static In, Dynamic Out: Counterfactual Action Augmentation for Moving Object Manipulation
- Georgia Institute of Technology（佐治亚理工学院）
机构由 AI 辅助整理，请以论文原文为准。
AI总结：
该研究针对视觉运动策略无法适应运动物体操作的问题，提出SIDO反事实动作增强方法，经模拟和真实任务验证，其在提升运动物体操作成功率的同时保留了静态物体操作性能。
AI中文摘要：
视觉运动策略在目标物体执行时保持静态的操作任务上已取得进展，但实际部署中该假设不成立：传送带上的部件会漂移，风中的水果会晃动。我们提出Static In, Dynamic Out（SIDO），一种反事实动作增强方法，使仅在静态物体演示上训练的策略能在测试时适应未见过的物体运动。核心思路是将运动物体操作分解为两个子问题：预测物体将到达的位置，以及到达该预测位姿。SIDO将物体位移到反事实未来位置，并调整演示的动作块以保持手-物体相对位姿，从而生成目标条件策略。部署时，物体位姿预测器提供未来位置。在三个模拟任务（Mug、Square、Stack）下的五种物体运动模式和两个真实世界任务（Gantry、Peachtree）中，SIDO在保持静态物体性能的同时，提升了运动物体操作的成功率。项目网站：this https URL。
英文摘要：
Visuomotor policies have advanced on manipulation tasks where the target object stays static during execution, but real deployments break this assumption: parts drift on conveyors and fruits sway in the wind. We introduce Static In, Dynamic Out (SIDO), a counterfactual action augmentation that enables a policy trained only on static object demonstrations to adapt to unseen object motion at test time. Our key idea is to factorize moving object manipulation into two sub-problems: predicting where the object will be, and reaching that predicted pose. SIDO displaces the object to a counterfactual future position and morphs the demonstrated action chunk to preserve the hand-object relative pose, yielding a goal-conditioned policy. At deployment an object pose predictor supplies the future position. Across three simulated tasks (Mug, Square, Stack) under five object motion patterns and two real-world tasks (Gantry, Peachtree), SIDO improves moving object success over the baselines while preserving static object performance. Project website: https://sido-staticindynamicout.github.io/.
