Bringing clothing into desired configurations with limited perception

Abstract

Fig. 1: The PR2 with a pair of pants in a crumpled initial configuration.Abstract-We consider the problem of autonomously bringing an article of clothing into a desired configuration using a general-purpose two-armed robot. We propose a hidden Markov model (HMM) for estimating the identity of the article and tracking the article's configuration throughout a specific sequence of manipulations and observations. At the end of this sequence, the article's configuration is known, though not necessarily desired. The estimated identity and configuration of the article are then used to plan a second sequence of manipulations that brings the article into the desired configuration. We propose a relaxation of a strainlimiting finite element model for cloth simulation that can be solved via convex optimization; this serves as the basis of the transition and observation models of the HMM. The observation model uses simple perceptual cues consisting of the height of the article when held by a single gripper and the silhouette of the article when held by two grippers. The model accurately estimates the identity and configuration of clothing articles, enabling our procedure to autonomously bring a variety of articles into desired configurations that are useful for other tasks, such as folding.

A 3D descriptor to detect task-oriented grasping points in clothing

“…The coarse-to-fine approach shows promising results dealing with the very challenging task of detecting task-oriented grasping points in garments, where the appearance is highly variable, and methods typically resort to long manipulation sequences to guarantee that the garment has been brought to a certain state, such as in [7].…”

“…In recent years Maitin-Shepard et al [6] and Cusumano-towner et al [7] demonstrated functional end-to-end systems for automatically handling clothes (albeit in very controlled settings that simplify perception). The proposed systems are able to pick up a piece of laundry and manipulate it until a desired configuration is reached.…”

Coupling cloth and rigid bodies for dexterous manipulation

“…Many previous approaches enhance control and planning algorithms by using simulation techniques to estimate the state of the deformable objects. For example, robotics researchers used a cloth simulator to approximate the contour of the clothes being folded by a PR2 robot [Cusumano-Towner et al 2011]. Because the interaction between the robot hands and the cloth was relatively simple (i.e.…”

POMDP approach to robotized clothes separation

“…Willimon et al [3] propose a system where the grasping operation is repeated as many times as necessary to ensure a correct grasp, and then four basic visual features are extracted to classify a previously learned object. Cusumano et al [4] use a robot to bring a piece of clothing to a specific configuration using predefined grasping actions by an arbitrary edge, and a series of lowhanging-point re-grasps.…”

“…Specially, removing more than one piece of clothing at once is a failure. Note that separation is a necessary previous step on a variety of tasks, like identification [1], [3] and folding [4]. …”