Treffer: A Framework for Digital Twin-based Robotic Cloth Manipulation.

Significant advancements have been made towards the automation of tasks involving highly deformable object manipulation. However, due to its complexity, predicting the behaviour of such objects, tasks requiring precision and flexibility remain far from full automation. In this paper, a method for the generation of a digital twin is proposed to serve as a solution for robotized cloth manipulation tasks. To estimate mechanical parameters and reconstruct a cloth’s digital twin, our approach integrates image processing and a genetic algorithm process. Specifically, a spatio-temporal graph of the cloth is built using Scale-Invariant Feature Transform (SIFT), and a genetic algorithm is used to optimize the parameters of a mass-spring-damper (MSD) model. Contrary to prior methods, which were relying on predefined cloth models or computationally expensive simulations (e.g., FEM), the proposed framework enables lightweight, data-driven parameter tuning from a single monocular RGB video showing the deformations of the cloth during a set of predefined movements by a pair of robotic arms. The generated digital twin is then used to calculate the optimal trajectory for the robotic arms in typical cloth manipulation tasks and is subsequently evaluated in real-world experiments. Tasks such as laying a cloth on a table and folding it in half are used to validate the method’s accuracy and applicability in both simulated and physical environments. The proposed method has potential applications in industrial textile handling, domestic service robotics, and elderly care, particularly in settings where low-cost, sensor-efficient automation is desirable. [ABSTRACT FROM AUTHOR]

Copyright of Journal of Intelligent & Robotic Systems is the property of Springer Nature and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)