Cloth Sim2Real Benchmark

Cloth Sim2Real Benchmark数据集由阿尔托大学和工业机器人与信息研究所合作创建，包含三块不同材质的布料（毛巾、亚麻和格子布），每块布料尺寸为50×70厘米。数据集通过动态和准静态布料操作任务收集，旨在评估模拟环境与现实世界之间的差距，特别是针对布料操作任务。该数据集用于评估四种流行的可变形物体模拟器：MuJoCo、Bullet、Flex和SOFA的性能。数据集的应用领域包括机器人布料操作技能的开发和评估，旨在解决模拟与现实之间的差距问题，提高机器人操作布料的准确性和效率。

The Cloth Sim2Real Benchmark dataset was co-developed by Aalto University and the Institute of Industrial Robotics and Information. It includes three types of cloth with distinct materials: towel, linen, and checkered cloth, each measuring 50 × 70 centimeters. The dataset is collected through dynamic and quasi-static cloth manipulation tasks, aiming to evaluate the sim-to-real gap specifically for cloth manipulation tasks. It is used to assess the performance of four mainstream deformable object simulators: MuJoCo, Bullet, Flex, and SOFA. The application domains of this dataset cover the development and evaluation of robotic cloth manipulation skills, with the objective of bridging the sim-to-real gap and improving the accuracy and efficiency of robotic cloth manipulation.