Learning of grasp selection based on shape-templates

By Alexander Herzog, Peter Pastor, Mrinal Kalakrishnan, Ludovic Righetti, Jeannette Bohg, Tamim Asfour, Stefan Schaal. 2013

134 citations recorded by the source corpus.

Subfields: Biomedical Engineering, Control and Systems Engineering

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Biomedical Engineering

In Control and Systems Engineering

Citations in the corpus

Learning hand-eye coordination for robotic grasping with deep learning and large-scale data collection (2016, methodology): “This stands in contrast to many successful prior grasping methods, many of which use large-scale simulated data to generate synthetic multiviewpoint depth images (Bohg et al., 2014; Gualtieri et al., 2016; Herzog et al., 2014), detect grasp affordance based on depth images (Lenz et al.”
Deep Reinforcement Learning for Vision-Based Robotic Grasping: A Simulated Comparative Evaluation of Off-Policy Methods (2018, methodology): “Prior methods have typically relied on one of three sources of supervision: human labels [17], [22], geometric criteria for grasp success computed offline [12], and robot self-supervision, measuring grasp success using sensors on the robot’s gripper [33].”
Recent Advances in Robot Learning from Demonstration (2020, unknown): “A popular category of policy learning methods assume direct access to the state and utilize it as the input to the policy (15, 16, 26, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84).”
One-shot learning and generation of dexterous grasps for novel objects (2016, background): “This problem can be overcome by letting the robot learn the mapping from vision to action (Bard and Troccaz, 1990; Coelho et al., 2000; Detry et al., 2013; Herzog et al., 2014; Kamon et al., 1996; Kim, 2007; Kroemer et al., 2010, 2012; Morales et al., 2004; Platt et al., 2006; Saxena et al., 2008; Z”
A Review of Robot Learning for Manipulation: Challenges, Representations, and Algorithms (2019, background): “To obtain valid grasp frames, the robot can learn a classifier to determine if sampled grasp frames will lead to successful grasps (Saxena et al., 2008; Nguyen and Kemp, 2014; Herzog et al., 2014; Pinto and Gupta, 2016), or sample from a learned probability density over successful grasps or cached g”
Grasp Pose Detection in Point Clouds (2017, methodology): “First, whereas the Herzog/Kappler method involves segmenting the object from the background and taking a convex hull or a bounding box around the Prepared using sagej.cls object (see the section on Grasp Heightmaps in Herzog et al. (2014)), our method does not.”
Dex-Net 1.0: A cloud-based network of 3D objects for robust grasp planning using a Multi-Armed Bandit model with correlated rewards (2016, methodology): “Another line of research has focused on synthesizing grasps using statistical models [5] learned from a database of images [27] or point clouds [9], [15], [46] of objects annotated with grasps from human demonstrators [15], [27] or physical execution [15].”
Learning Hand-Eye Coordination for Robotic Grasping with Large-Scale Data Collection (2016, background): “Such methods take a variety of forms, including human-supervised methods that predict grasp configurations [Herzog et al., 2014, Lenz et al., 2015] and methods that predict finger placement from geometric criteria computed offline [Goldfeder et al., 2009].”
Action Image Representation: Learning Scalable Deep Grasping Policies with Zero Real World Data (2020, background): “Recently, data-driven methods have become popular [3], due to their flexibility in specifying sensory input and ease of incorporating data from experience through real robot grasp executions [9].”
Leveraging big data for grasp planning (2015, methodology): “It is a modification of the features used in [22] and does not require knowledge about object identity, category, pose or accurate segmentation.”