Learning to Fold Real Garments with One Arm: A Case Study in Cloud-Based Robotics Research

TLDR

Autonomous fabric manipulation remains a challenging robotics problem, hindered by high hardware costs and diverse platforms that make progress evaluation difficult. The study benchmarks fabric manipulation algorithms on physical robots using the Reach cloud robotics platform. Four novel learning‑based algorithms were developed to model expert actions, keypoints, reward functions, and dynamic motions, and compared against four learning‑free and inverse dynamics algorithms for folding a crumpled T‑shirt with a single arm, with data collection, training, and evaluation performed remotely via Reach. The hybrid imitation‑learning algorithm achieved human‑level performance on flattening and 93 % of human performance on folding. All data, code, models, and supplemental material are available at https://sites.google.com/berkeley.edu/cloudfolding.

Abstract

Autonomous fabric manipulation is a longstanding challenge in robotics, but evaluating progress is difficult due to the cost and diversity of robot hardware. Using Reach, a cloud robotics platform that enables low-latency remote execution of control policies on physical robots, we present the first systematic benchmarking of fabric manipulation al-gorithms on physical hardware. We develop 4 novel learning-based algorithms that model expert actions, keypoints, reward functions, and dynamic motions, and we compare these against 4 learning-free and inverse dynamics algorithms on the task of folding a crumpled T-shirt with a single robot arm. The entire lifecycle of data collection, model training, and policy evaluation was performed remotely without physical access to the robot workcell. Results suggest a new algorithm combining imitation learning with analytic methods achieves human-level performance on the flattening task and 93% of human-level performance on the folding task. See https://sites.google.com/berkeley.edu/ cloudfolding for all data, code, models, and supplemental material.