FolDeX provides real-robot benchmark for long deformable object manipulation
FolDeX: A Physical-World Benchmark for Long-Horizon Robotic Manipulation of Deformable Objects
Robotics
Summary
Robots often struggle to handle soft, flexible objects in the real world, especially when tasks require many steps. The authors introduce FolDeX, a large collection of real robot data focused on folding clothes, which is a challenging long task involving soft items. This benchmark helps test and improve robot skills by allowing researchers to reuse data from different tasks, objects, and setups. It includes over 2,000 hours of robotic activity and a public evaluation platform for testing robot policies on consistent tasks and conditions.
What this means in practice
- •For robotics engineers: Develop and test reliable long-step robot programs for manipulating soft items using a large real-world garment folding dataset and benchmark.
- •For industrial automation teams: Improve factory robotic systems handling flexible products by using FolDeX to train and evaluate multi-stage manipulation under real conditions.
Authors
Chenhuan Liu, Yi Xu, Feng Wu, Hanyang Wang, Wenxiao Kuai, Weihao Ding, Shan Wang, Yang Liu, Shuyong Gao, Wenqiang Zhang
Abstract
Embodied AI, including vision-language-action and world-action models, must operate reliably in the physical world. Yet methods that perform well in simulation can degrade substantially on real robots, especially in long-horizon deformable-object manipulation, where policies must track changing states and execute reliable multi-stage bimanual interactions. Existing real-robot benchmarks mainly focus on short-horizon rigid-object tasks and offer limited coverage of long-horizon deformable manipulation. We introduce FolDeX, a physical-world benchmark built entirely from real-robot data, with garment folding as its primary task. Since real-robot data collection is costly, FolDeX studies how heterogeneous physical experience can be reused efficiently. The benchmark is organized around four research axes: leveraging human intervention and recovery data collected during deployment; transferring data across tasks, including across garment categories and from rigid to deformable-object manipulation; reusing data across scenes with changes in lighting, background, and layout; and transferring data across robotic embodiments. FolDeX provides 2,000+ hours of real-robot data spanning 20+ tasks and 10+ embodiments. We also establish a fair real-robot evaluation platform for externally submitted policies, with standardized tasks, held-out physical objects, controlled initializations, and a unified execution protocol. The platform is publicly accessible at https://ai.midea.com/#/fold-challenge. We hope FolDeX will serve as a unified testbed for heterogeneous real-robot data reuse and reliable long-horizon deformable manipulation.