Generalization of human grasping for multi-fingered robot hands

Amor, Heni Ben; Kroemer, Oliver; Hillenbrand, Ulrich; Neumann, Gerhard; Peters, Jan

doi:10.1109/iros.2012.6386072

Cited by 57 publications

(36 citation statements)

References 29 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A known object's geometry is warped until it matches that of a novel object, thereby also warping grasp points on the surface of the known object onto candidate grasp points on the novel object. Ben Amor et al (2012) exploit this warping method to transfer grasps taught by a human hand (using a data glove) to contact points for a robot hand on a novel object. We compute a full hand grasping configuration for a novel object, using a grasp model that is learned from a single or a few example grasps.…”

Section: Ijrr -(-)mentioning

confidence: 99%

“…This works well for low DoF hands. Another class of approaches captures the global properties of the hand shape either at the point of grasping, or during the approach (Ben Amor et al, 2012). This global hand shape can additionally be associated with global object shape, allowing generalisation by warping grasps to match warps of global object shape (Hillenbrand & Roa, 2012).…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

One-shot learning and generation of dexterous grasps for novel objects

Kopicki

Detry

Adjigble

et al. 2015

The International Journal of Robotics Research

117

163

View full text Add to dashboard Cite

This paper presents a method for one-shot learning of dexterous grasps, and grasp generation for novel objects. A model of each grasp type is learned from a single kinesthetic demonstration, and several types are taught. These models are used to select and generate grasps for unfamiliar objects. Both the learning and generation stages use an incomplete point cloud from a depth camera -no prior model of object shape is used. The learned model is a product of experts, in which experts are of two types. The first is a contact model and is a density over the pose of a single hand link relative to the local object surface. The second is the hand configuration model and is a density over the whole hand configuration. Grasp generation for an unfamiliar object optimises the product of these two model types, generating thousands of grasp candidates in under 30 seconds. The method is robust to incomplete data at both training and testing stages. When several grasp types are considered the method selects the highest likelihood grasp across all the types. In an experiment, the training set consisted of five different grasps, and the test set of forty-five previously unseen objects. The success rate of the first choice grasp is 84.4% or 77.7% if seven views or a single view of the test object are taken, respectively. Keywords learning, dexterous grasping * Authors Kopicki, Detry and Wyatt are identified as the primary authors of this work. Kopicki is identified as the first author. † Corresponding author; e-mail: jlw@cs.bham.ac.uk 2 IJRR -(-) Fig. 1. Leftmost image: Objects used, the four objects on the left were used solely for training, the remaining forty three objects on the right were solely used as novel test objects. Rightmost image: The Boris manipulation platform on which the experiments reported were carried out.

show abstract

Section: Ijrr -(-)mentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

One-shot learning and generation of dexterous grasps for novel objects

Kopicki

Detry

Adjigble

et al. 2015

The International Journal of Robotics Research

117

163

View full text Add to dashboard Cite

show abstract

“…A subclass of these focused on learning a mapping from incomplete object views to grasp parameters [1], [8], [11], [15], [14], [19], while another subclass has aimed at transferring grasps across objects whose complete 3D shape is known [2], [12]. When addressing the problem of grasping a partiallyperceived object, authors have developed means of learning how to place the wrist of the gripper with respect to the part of the object that is perceived by the robot's vision system [8], [11], [15], [14].…”

Section: Related Workmentioning

confidence: 99%

“…This works well for low DoF hands. Another class of approaches captures the global properties of the hand shape either at the point of grasping, or during the approach [2]. This global hand shape can additionally be associated with global object shape, allowing generalisation by warping grasps to match warps of global object shape [12].…”

Section: Introductionmentioning

confidence: 99%

Learning dexterous grasps that generalise to novel objects by combining hand and contact models

Kopicki

Detry

Schmidt

et al. 2014

2014 IEEE International Conference on Robotics and Automation (ICRA)

View full text Add to dashboard Cite

Abstract-Generalising dexterous grasps to novel objects is an open problem. We show how to learn grasps for high DoF hands that generalise to novel objects, given as little as one demonstrated grasp. During grasp learning two types of probability density are learned that model the demonstrated grasp. The first density type (the contact model) models the relationship of an individual finger part to local surface features at its contact point. The second density type (the hand configuration model) models the whole hand configuration during the approach to grasp. When presented with a new object, many candidate grasps are generated, and a kinematically feasible grasp is selected that maximises the product of these densities. We demonstrate 31 successful grasps on novel objects (an 86% success rate), transferred from 16 training grasps. The method enables: transfer of dexterous grasps within object categories; across object categories; to and from objects where there is no complete model of the object available; and using two different dexterous hands.

show abstract

“…Santello et al [10] show that the variation in final imagined grasp poses for a large number of objects is quite small, with well over 80% of the variation explained by only two principal components. Recently, Ben Amor et al [11] used a low-dimensional sub-space built from a database of recorded human grasping postures to perform grasp optimization on a robot. The focus of their work is complementary to ours in that their approach synthesizes motions for "reach-and-grasp" tasks.…”

Section: Related Workmentioning

confidence: 99%

Goal directed multi-finger manipulation: Control policies and analysis

Andrews

Kry

2013

Computers & Graphics

View full text Add to dashboard Cite

We present a method for one-handed, task-based manipulation of objects. Our approach uses a mid-level, multi-phase approach to organize the problem into three phases. This provides an appropriate control strategy for each phase and results in cyclic finger motions that, together, accomplish the task. The exact trajectory of the object is never specified since the goal is defined by the final orientation and position of the object. All motion is physically based and guided by a control policy that is learned through a series of offline simulations. We also discuss practical considerations for our learning method. Variations in the synthesized motions are possible by tuning a scalarized multi-objective optimization. We demonstrate our method with two manipulation tasks, discussing the performance and limitations. Additionally, we provide an analysis of the robustness of the low-level controllers used by our framework.

show abstract

Generalization of human grasping for multi-fingered robot hands

Cited by 57 publications

References 29 publications

One-shot learning and generation of dexterous grasps for novel objects

One-shot learning and generation of dexterous grasps for novel objects

Learning dexterous grasps that generalise to novel objects by combining hand and contact models

Goal directed multi-finger manipulation: Control policies and analysis

Contact Info

Product

Resources

About