BoxTwin: Learning Elastoplastic Articulated Object Dynamics from Videos
For robotics, this enables predictive and adaptive control of deformable articulated objects in unstructured environments, addressing a gap in existing digital twin models.
BoxTwin learns the full dynamics of elastoplastic articulated objects from videos, accurately tracking joint trajectories and reproducing post-contact plastic behavior over long horizons in manual folding and dual-arm manipulation experiments.
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.