外观
论文arXiv 2608.227011 个月前
物理滤波促进机器人学习的泛化
原文链接：Physics filtering favors the generalization of robot learning（arXiv 2608.22701 · 发布日期：2026-09-04）
Physics filtering favors the generalization of robot learning
论文提出PhyFilter，用物理滤波反馈残差在有限数据下校正学习输出，并在四类机器人系统上验证泛化。
核心要点
- 动机：大规模采集真实示范成本高且缓慢，机器人需可靠泛化到未见环境。
- 方法：PhyFilter以物理滤波的学习残差校正策略输出，轻量且模型无关。
- 其参数可由自动学习算法优化，避免手动调参并便于接入多种机器人策略。
- 实验显示四足、无人机、空中机械臂和加速度微分器在分布变化下更稳健。
验证范围与局限：验证范围为四类代表性机器人系统及未见地形、载荷、速度、风扰、质量不确定性和分布偏移；摘要未说明仿真或真机设置。
英文摘要 (Abstract)
Abstract Living organisms exhibit extraordinary adaptability to unseen environments through their intrinsic physical structures and lifelong feedback-driven learning. Endowing robots with comparable generalization is critical for reliable operation in the real world. While recent approaches attempt to improve generalization by scaling training data, such strategies remain impractical for robotics, where collecting real-world demonstrations at the scale of large language models is prohibitively costly and slow. Contrary to this reliance on massive datasets, we show that robots can generalize effectively even with limited training data by leveraging a feedback mechanism, namely PhyFilter, that corrects learning outputs with physics-filtered learning residuals. PhyFilter operates as a lightweight, model-agnostic module whose parameters can be automatically optimized through an auto-learning algorithm, eliminating manual tuning and enabling seamless integration with diverse robot policies. We validate PhyFilter across four representative robotic systems, demonstrating that it enables quadruped robots to generalize to unseen terrains, payload variations, and speed ranges; drones to flight under unseen wind disturbances; aerial manipulators to achieve centimeter-level in-air capture despite wind and mass uncertainties; and acceleration differentiators to remain robust with distribution shift. These results reveal a scalable and data-efficient pathway toward generalizable machine intelligence, showing that physics-filtered feedback can serve as a powerful alternative to massive data scaling.
作者：Jindou Jia, Shixuan Han, M M Wang, Gen Li, Zihan Yang, Sicheng Zhou, Kexin Guo, Jianfei Yang, Xiang Yu, Wei Wang
来源
本页事实对应的公开页面，以原文为准。