BoxTwin：从视频中学习弹塑性铰接物体动力学
BoxTwin: Learning Elastoplastic Articulated Object Dynamics from Videos
- SceniX, Inc.（SceniX公司）
- University of British Columbia（英属哥伦比亚大学）
机构由 AI 辅助整理，请以论文原文为准。
AI总结：
研究针对现有模型处理弹塑性铰接物体的不足，提出BoxTwin框架，通过视频学习其动力学，经重建场景、识别本构模型等流程，实验证明该框架能准确跟踪轨迹并再现塑性行为，推动数字孪生用于可变形铰接物体的控制。
AI中文摘要：
数字孪生使机器人能够预测并适应物理交互，但现有模型难以处理具有非线性弹性、塑性屈服和损伤累积的弹塑性铰接物体（EAO）。我们提出了BoxTwin，一个交互式数字孪生框架，可从视频中学习EAO的完整动力学。我们的流程重建场景，为每个EAO识别一个物理感知本构模型。在EAO的手动折叠和双臂操作实验表明，BoxTwin能准确跟踪关节轨迹并在长时间内再现接触后的塑性行为。通过将视频驱动重建与弹塑性损伤建模相结合，BoxTwin推动数字孪生向非结构化环境中可变形铰接物体的预测性、自适应控制发展。
英文摘要：
Digital twins enable robots to anticipate and adapt to physical interactions, but existing models struggle with elastoplastic articulated objects (EAOs) that exhibit nonlinear elasticity, plastic yielding, and damage accumulation. We present BoxTwin, an interactive digital twin framework that learns the full dynamics of EAOs from videos. Our pipeline reconstructs the scene, identifies a physics aware constitutive model for each EAO. Experiments on manual folding and dual arm manipulation of EAOs show that BoxTwin accurately tracks joint trajectories and reproduces post contact plastic behavior over long horizons. By integrating video driven reconstruction with elastoplastic damage modeling, BoxTwin advances digital twins toward predictive, adaptive control of deformable articulated objects in unstructured environments.
