FBI：通过动态视觉触觉捷径策略学习灵巧的手内操作
FBI: Learning Dexterous In-hand Manipulation with Dynamic Visuotactile Shortcut Policy
AI总结：
针对机器人灵巧手内操作因模态单一限制适应性的问题，提出FBI视觉触觉模仿学习框架，通过动态融合触觉与视觉观测，建立触觉信号与物体运动的因果联系，在多任务中优于基线方法。
摘要：
由于复杂的接触动力学和部分可观测性，灵巧的手内操作一直是机器人技术中的长期挑战。虽然人类在这类任务中协同使用视觉和触觉，但机器人方法往往优先考虑一种模态，因此限制了适应性。本文介绍了Flow Before Imitation（FBI），这是一种视觉触觉模仿学习框架，通过运动动力学将触觉交互与视觉观测动态融合。与之前的静态融合方法不同，FBI通过动力学感知的潜在模型在触觉信号和物体运动之间建立因果联系。FBI采用基于Transformer的交互模块，将流动衍生的触觉特征与视觉输入融合，训练单步扩散策略以实现实时执行。大量实验表明，该方法在两个定制的手内操作任务和三个标准灵巧操作任务上，在仿真和现实世界中均优于基线方法。代码、模型和更多结果可在网站https://sites.google.com/view/dex-fbi获取。
英文摘要：
Dexterous in-hand manipulation is a long-standing challenge in robotics due to complex contact dynamics and partial observability. While humans synergize vision and touch for such tasks, robotic approaches often prioritize one modality, therefore limiting adaptability. This paper introduces Flow Before Imitation (FBI), a visuotactile imitation learning framework that dynamically fuses tactile interactions with visual observations through motion dynamics. Unlike prior static fusion methods, FBI establishes a causal link between tactile signals and object motion via a dynamics-aware latent model. FBI employs a transformer-based interaction module to fuse flow-derived tactile features with visual inputs, training a one-step diffusion policy for real-time execution. Extensive experiments demonstrate that the proposed method outperforms the baseline methods in both simulation and the real world on two customized in-hand manipulation tasks and three standard dexterous manipulation tasks. Code, models, and more results are available in the website https://sites.google.com/view/dex-fbi.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
