EGAD! An Evolved Grasping Analysis Dataset for Diversity and Reproducibility in Robotic Manipulation

Morrison, Douglas; Corke, Peter; Leitner, Jürgen

doi:10.1109/lra.2020.2992195

Cited by 104 publications

(82 citation statements)

References 43 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We generated one training set of cluttered scenes on 18 objects from the YCB object set [37] and 120 objects from the EGAD! training set [38]. We also generated three validation sets: one on the 49 validation objects in the EGAD!…”

Section: Ddgc Datasetmentioning

confidence: 99%

DDGC: Generative Deep Dexterous Grasping in Clutter

Lundell

Verdoja

Kyrki

2021

IEEE Robot. Autom. Lett.

View full text Add to dashboard Cite

Recent advances in multi-fingered robotic grasping have enabled fast 6-Degrees-of-Freedom (DOF) single object grasping. Multi-finger grasping in cluttered scenes, on the other hand, remains mostly unexplored due to the added difficulty of reasoning over obstacles which greatly increases the computational time to generate high-quality collision-free grasps. In this work, we address such limitations by introducing DDGC, a fast generative multifinger grasp sampling method that can generate high quality grasps in cluttered scenes from a single RGB-D image. DDGC is built as a network that encodes scene information to produce coarse-to-fine collision-free grasp poses and configurations. We experimentally benchmark DDGC against two state-of-the-art methods on 1200 simulated cluttered scenes and 7 real-world scenes. The results show that DDGC outperforms the baselines in synthesizing highquality grasps and removing clutter. DDGC is also 4-5 times faster than GraspIt!. This, in turn, opens the door for using multi-finger grasps in practical applications which has so far been limited due to the excessive computation time needed by other methods. Code and videos are available at https://irobotics.aalto.fi/ddgc/.

show abstract

Section: Ddgc Datasetmentioning

confidence: 99%

DDGC: Generative Deep Dexterous Grasping in Clutter

Lundell

Verdoja

Kyrki

2021

IEEE Robot. Autom. Lett.

View full text Add to dashboard Cite

show abstract

“…Object Datasets. To get a diverse set of objects that demand different grasp strategies, we use the objects from the recently introduced EGAD dataset [45] and select 48 objects of varying grasp difficulties to assist evaluating the performance of the optimized hand on power and pinch These objects are either selected from the EGAD dataset [45] or procedurally generated.…”

Section: A Experiments Setupmentioning

confidence: 99%

“…Fig. 4: (a) Three main grasp types of the human grasp taxonomy: power grasp, lateral grasp, and pinch grasp (images are taken from Feix et al [11]); (b) Example objects we used for evaluating grasps.These objects are either selected from the EGAD dataset[45] or procedurally generated.…”

mentioning

confidence: 99%

Emergent Hand Morphology and Control from Optimizing Robust Grasps of Diverse Objects

Pan

Garg

Anandkumar

et al. 2021

2021 IEEE International Conference on Robotics and Automation (ICRA)

View full text Add to dashboard Cite

Evolution in nature illustrates that the creatures' biological structure and their sensorimotor skills adapt to the environmental changes for survival. Likewise, the ability to morph and acquire new skills can facilitate an embodied agent to solve tasks of varying complexities. In this work, we introduce a data-driven approach where effective hand designs naturally emerge for the purpose of grasping diverse objects. Jointly optimizing morphology and control imposes computational challenges since it requires constant evaluation of a black-box function that measures the performance of a combination of embodiment and behavior. We develop a novel Bayesian Optimization algorithm that efficiently co-designs the morphology and grasping skills jointly through learned latentspace representations. We design the grasping tasks based on a taxonomy of three human grasp types: power grasp, pinch grasp, and lateral grasp. Through experimentation and comparative study, we demonstrate the effectiveness of our approach in discovering robust and cost-efficient hand morphologies for grasping novel objects. https://xinleipan.github.io/ emergent_morphology/ 1 NVIDIA,

show abstract

“…This has led to a hand-centric design paradigm (2) where robotic grippers take the form of an articulated set of locally-controlled rigid links, while also relying on an opto-motor feedback loop linking perception, planning, and action to achieve a grasping goal. Modern rigid grippers show great promise with many controllable degrees of freedom and embedded sensors (5)(6)(7)(8)(9), but can present challenges for grasp planning and control in the presence of uncertainty, or with complex target geometries (10,11).…”

mentioning

confidence: 99%

“…Our empirical investigation of grasping performance using non-deterministic entanglement is particularly successful in grasping topologically and geometrically complex objects (10,11) without the need for planning, but has trouble with simpler objects like spheres and vertical tubes where traditional deterministic grippers work well, e.g., the YCB object set of generally cylindrical, spherical, and cuboidal targets (35).…”

mentioning

confidence: 99%

Grasping via entanglement

Wood

Becker

Teeple

et al. 2021

Preprint

View full text Add to dashboard Cite

We present a grasping strategy for complex shapes via the collective entanglement of and by an array of actuated filaments. The basic unit of this array is a slender hollow elastomeric filament that is pneumatically actuated to form a highly curved structure. The multiple self and mutual contact interactions between the filaments and the object create a randomly tangled spatial assemblage that enables a soft conformable grasp. We realize this using a gripper enabled by a new fabrication method to create inexpensive and modular arrays of fluidically actuated elastomeric filaments. We demonstrate that a collective of highly compliant filamentous actuators is capable of a soft, adaptable grasp across a range of loads that vary in size, shape, and geometric and topological complexity without any feedback. A theoretical framework for the collective mechanics of filaments in contact with complex objects allows us to explain our experimental findings, while a phase diagram characterizes the design space in terms of the properties of the gripper and the target. Overall, our grasping approach adapts to the mechanical, geometric, and topological complexity of target objects via an uncontrolled, spatially distributed, and heterogeneous scheme without perception or planning, in sharp contrast with current deterministic feedback-driven robotic grasping methods.

show abstract

EGAD! An Evolved Grasping Analysis Dataset for Diversity and Reproducibility in Robotic Manipulation

Cited by 104 publications

References 43 publications

DDGC: Generative Deep Dexterous Grasping in Clutter

DDGC: Generative Deep Dexterous Grasping in Clutter

Emergent Hand Morphology and Control from Optimizing Robust Grasps of Diverse Objects

Grasping via entanglement

Contact Info

Product

Resources

About