Data-Driven Grasp Synthesis—A Survey
By Jeannette Bohg, Antonio Morales, Tamim Asfour, Danica Kragic. 2013
1,118 citations recorded by the source corpus.
We review the work on data-driven grasp synthesis and the methodologies for sampling and ranking candidate grasps. We divide the approaches into three groups based on whether they synthesize grasps for known, familiar, or unknown objects. This structure allows us to identify common object representations and perceptual processes that facilitate the employed data-driven grasp synthesis technique. In the case of known objects, we concentrate on the approaches that are based on object recognition and pose estimation. In the case of familiar objects, the techniques use some form of a similarity matching to a set of previously encountered objects. Finally, for the approaches dealing with unknown objects, the core part is the extraction of specific features that are indicative of good grasps. Our survey provides an overview of the different methodologies and discusses open problems in the area of robot grasping. We also draw a parallel to the classical approaches that rely on analytic formulations.
Subfields: Cognitive Neuroscience, Control and Systems Engineering, Robotics & Control
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Cognitive Neuroscience
- Sergey Levine
- Chelsea Finn
- Pieter Abbeel
- Jan Peters
- Gerhard Neumann
- Oliver Kroemer
- J. Bagnell
- Vladlen Koltun
In Control and Systems Engineering
- Sergey Levine
- Pieter Abbeel
- Chelsea Finn
- Abhinav Kumar Gupta
- Dieter Fox
- Lerrel Pinto
- Vikash Kumar
- Jan Peters
In Robotics & Control
- Sergey Levine
- Pieter Abbeel
- Chelsea Finn
- Abhinav Kumar Gupta
- Dieter Fox
- Lerrel Pinto
- Jan Peters
- Silvio Savarese
Citations in the corpus
- Habitat 2.0: Training Home Assistants to Rearrange their Habitat (2021, background): “[24] Jeannette Bohg, Antonio Morales, Tamim Asfour, and Danica Kragic.”
- QT-Opt: Scalable Deep Reinforcement Learning for Vision-Based Robotic Manipulation (2018, background): “This stands in contrast to the kinds of grasping behaviors observed in humans and animals, where the grasp is a dynamical process that tightly interleaves sensing and control at every stage [9, 10].”
- Learning hand-eye coordination for robotic grasping with deep learning and large-scale data collection (2016, background): “While a complete survey of grasping is outside the scope of this work, we refer the reader to standard surveys on the subject for a more complete treatment (Bohg et al., 2014).”
- Deep learning for detecting robotic grasps (2013, methodology): “We use the X-Y coordinates of the rectangle center along with the depth of the closest point to determine a grasping point in the robot’s coordinate frame.”
- Using Simulation and Domain Adaptation to Improve Efficiency of Deep Robotic Grasping (2018, background): “While a complete survey of grasping is outside the scope of this work, we refer the reader to standard surveys on the subject for a more complete treatment [2].”
- The Curious Robot: Learning Visual Representations via Physical Interactions (2016, background): “A comprehensive literature review of this area can be found in [17,18].”
- Dex-Net 2.0: Deep Learning to Plan Robust Grasps with Synthetic Point Clouds and Analytic Grasp Metrics (2017, background): “known objects to models in the database [4, 5, 8, 23, 27].”
- Supervision via competition: Robot adversaries for learning tasks (2016, neutral): “Refer to [32], [33] for surveys of prior work.”
- Learning to push by grasping: Using multiple tasks for effective learning (2016, background): “For a comprehensive literature review, we direct the readers to [8], [9].”
- Robot Learning in Homes: Improving Generalization and Reducing Dataset Bias (2018, methodology): “Grasping is one of the fundamental problems in robotic manipulation and we refer readers to recent surveys Bicchi and Kumar [31], Bohg et al. [32] for a comprehensive review.”
- Sim-To-Real via Sim-To-Sim: Data-Efficient Robotic Grasping via Randomized-To-Canonical Adaptation Networks (2018, background): “Robotic grasping is a well studied problem [2].”
- Manipulation (2018, methodology): “How do we predict grasp locations for this object? One approach is to fit 3D models to these objects, or to use a 3D depth sensor, and perform analytical 3D reasoning to predict the grasp locations [1]–[4].”
- Trends and challenges in robot manipulation (2019, background): “Thus, many of the more recent approaches are data driven (37).”
- Leveraging big data for grasp planning (2015, methodology): “Data-driven methods towards grasping [10] address the challenges of grasp synthesis that arise in the real world such as noisy sensors and incomplete information about the object”
- ContactGrasp: Functional Multi-finger Grasp Synthesis from Contact (2019, background): “We refer the reader to [25] for an extensive survey of data-driven grasp synthesis.”
- PiGraphs (2016, background): “In particular, designing 3D scenes and virtual character interactions within them is still a time–consuming task requiring expertise and much manual effort.”
- BiGS: BioTac Grasp Stability Dataset (2016, background): “Even though there has been a lot of progress in the area of grasping, it is still considered an open challenge and even the state-of-the-art grasping methods may result in failures [1].”
- Deep Reinforcement Learning for Vision-Based Robotic Grasping: A Simulated Comparative Evaluation of Off-Policy Methods (2018, background): “Prior work in robotic manipulation has sought to address the grasping problem through a wide range of methods, from analytic grasp metrics [43], [36] to learning-based approaches [2].”