skip to main content


Title: Cloth Region Segmentation for Robust Grasp Selection
Cloth detection and manipulation is a common task in domestic and industrial settings, yet such tasks remain a challenge for robots due to cloth deformability. Furthermore, in many cloth-related tasks like laundry folding and bed making, it is crucial to manipulate specific regions like edges and corners, as opposed to folds. In this work, we focus on the problem of segmenting and grasping these key regions. Our approach trains a network to segment the edges and corners of a cloth from a depth image, distinguishing such regions from wrinkles or folds. We also provide a novel algorithm for estimating the grasp location, direction, and directional uncertainty from the segmentation. We demonstrate our method on a real robot system and show that it outperforms baseline methods on grasping success.  more » « less
Award ID(s):
1849154
NSF-PAR ID:
10210628
Author(s) / Creator(s):
; ; ; ;
Date Published:
Journal Name:
Proceedings of the IEEERSJ International Conference on Intelligent Robots and Systems
ISSN:
2153-0858
Format(s):
Medium: X
Sponsoring Org:
National Science Foundation
More Like this
  1. Robotic grasping is successful when a robot can sense and grasp an object without letting it slip. Beyond industrial robotic tasks, there are two main robotic grasping methods. The first is planning-based grasping where the object geometry is known beforehand and stable grasps are calculated using algorithms [1]. The second uses tactile feedback. Currently, there are capacitive sensors placed beneath stiff pads on the front of robotic fingers [2]. With post-execution grasp adjustment procedures to estimate grasp stability, a support vector machine classifier can distinguish stable and unstable grasps. The accuracy across the classes of tested objects is 81% [1]. We are proposing to improve the classifier's accuracy by wrapping flexible sensors around the robotic finger to gain information from the edges and sides of the finger. 
    more » « less
  2. null (Ed.)
    Silver nanocubes have found use in an array of applications but their performance has been plagued by the shape instability arising from the oxidation and dissolution of Ag atoms from the edges and corners. Here we demonstrate that the shape of Ag nanocubes can be well preserved by covering their edges and corners with a corrosion-resistant metal such as Ir. In a typical process, we titrate a Na 3 IrCl 6 solution in ethylene glycol (EG) into a suspension of Ag nanocubes in an EG solution in the presence of poly(vinylpyrrolidone) (PVP) held at 110 °C. The Ir atoms derived from the reduction of Na 3 IrCl 6 by EG and Ag are deposited onto the edges and then corners for the generation of Ag–Ir core-frame nanocubes. Remarkably, our results indicate that a small amount of Ir atoms on the edges and corners is adequate to prevent the Ag nanocubes from transforming into nanospheres when heated in a PVP/EG solution up to 110 °C. We further demonstrate that these Ag–Ir nanocubes embrace plasmonic properties comparable to those of the original Ag nanocubes, making them immediately useful in a variety of applications. This strategy for stabilizing the shape of Ag nanocubes should be extendible to Ag nanocrystals with other shapes or nanocrystals comprised of other metals. 
    more » « less
  3. We report a facile route to the synthesis of Ag@Au–Pt trimetallic nanocubes in which the Ag, Au, and Pt atoms are exposed at the corners, side faces, and edges, respectively. Our success relies on the use of Ag@Au nanocubes, with Ag 2 O patches at the corners and Au on the side faces and edges, as seeds for the site-selective deposition of Pt on the edges only in a reaction system containing ascorbic acid (H 2 Asc) and poly(vinylpyrrolidone). At an initial pH of 3.2, H 2 Asc can dissolve the Ag 2 O patches, exposing the Ag atoms at the corners of a nanocube. Upon the injection of the H 2 PtCl 6 precursor, the Pt atoms derived from the reduction by both H 2 Asc and Ag are preferentially deposited on the edges, leading to the formation of Ag@Au–Pt trimetallic nanocubes. We demonstrate the use of 2,6-dimethylphenyl isocyanide as a molecular probe to confirm and monitor the deposition of Pt atoms on the edges of nanocubes through surface-enhanced Raman scattering (SERS). We further explore the use of these bifunctional trimetallic nanoparticles with integrated plasmonic and catalytic properties for in situ SERS monitoring the reduction of 4-nitrothiophenol by NaBH 4 . Upon the removal of Ag via H 2 O 2 etching, the Ag@Au–Pt nanocubes evolve into trimetallic nanoboxes with a wall thickness of about 2 nm and well-defined openings at the corners. The trimetallic nanoboxes embrace plasmon resonance peaks in the near-infrared region with potential in biomedical applications. 
    more » « less
  4. null (Ed.)
    Phanerozoic limestones are composed of low-Mg calcite microcrystals (i.e., micrite) that typically measure between 1 and 9 lm in diameter. These microcrystals, which host most of the microporosity in subsurface reservoirs, are characterized by a variety of microtextures. Despite the overwhelming consensus that calcite microcrystals are diagenetic, the origin of the various textures is widely debated. The most commonly reported texture is characterized by polyhedral and rounded calcite microcrystals, which are interpreted to form via partial dissolution of rhombic microcrystals during burial diagenesis. A proposed implication of this model is that dissolution during burial is responsible for significant porosity generation. This claim has been previously criticized based on mass balance considerations and geochemical constrains. To explicitly test the dissolution model, a series of laboratory experiments were conducted whereby various types of calcites composed of rhombic and polyhedral microcrystals were partially dissolved under a constant degree of undersaturation, both near and far-from-equilibrium. Our results indicate that calcite crystals dissolved under far-from-equilibrium conditions develop rounded edges and corners, inter-crystal gulfs (narrow grooves or channels between adjacent crystals), and a few etch pits on crystal faces—observations consistent with the burial-dissolution hypothesis. Crystals dissolved under near-equilibrium conditions, in contrast, retain sharp edges and corners and develop ledges and pits—suggesting that dissolution occurs more selectively at high-energy sites. These observations support the longstanding understanding that far-from equilibrium dissolution is transport-controlled, and near-equilibrium dissolution is surface-controlled. Our results also show that while the rhombic calcite crystals may develop rounded edges and corners when dissolved under far from- equilibrium conditions the crystals themselves do not become spherical. By contrast, polyhedral crystals not only develop rounded edges and corners when dissolved under far-from-equilibrium conditions but become nearly spherical with continued dissolution. Collectively, these observations suggest that rounded calcite microcrystals more likely form from a precursor exhibiting an equant polyhedral texture, rather than from a euhedral rhombic precursor as previously proposed. Lastly, the observation that calcite crystals developed rounded edges and corners and intercrystal gulfs after only 5%dissolution indicates that the presence of such features in natural limestones need not imply that significant porosity generation has occurred. 
    more » « less
  5. null (Ed.)
    Humans leverage the dynamics of the environment and their own bodies to accomplish challenging tasks such as grasping an object while walking past it or pushing off a wall to turn a corner. Such tasks often involve switching dynamics as the robot makes and breaks contact. Learning these dynamics is a challenging problem and prone to model inaccuracies, especially near contact regions. In this work, we present a framework for learning composite dynamical behaviors from expert demonstrations. We learn a switching linear dynamical model with contacts encoded in switching conditions as a close approximation of our system dynamics. We then use discrete-time LQR as the differentiable policy class for data-efficient learning of control to develop a control strategy that operates over multiple dynamical modes and takes into account discontinuities due to contact. In addition to predicting interactions with the environment, our policy effectively reacts to inaccurate predictions such as unanticipated contacts. Through simulation and real world experiments, we demonstrate generalization of learned behaviors to different scenarios and robustness to model inaccuracies during execution. 
    more » « less