Results 1 - 10
of
15
Learning grasp context distinctions that generalize
- In Proceedings of the IEEE-RAS International Conference on Humanoid Robots
, 2006
"... Abstract — Control-based approaches to grasp synthesis create grasping behavior by sequencing and combining control primitives. In the absence of any other structure, these approaches must evaluate a large number of feasible control sequences as a function of object shape, object pose, and task. Thi ..."
Abstract
-
Cited by 9 (0 self)
- Add to MetaCart
(Show Context)
Abstract — Control-based approaches to grasp synthesis create grasping behavior by sequencing and combining control primitives. In the absence of any other structure, these approaches must evaluate a large number of feasible control sequences as a function of object shape, object pose, and task. This paper explores a new approach to grasp synthesis that limits consideration to variations on a generalized localize-reach-grasp control policy. A new learning algorithm, known as schema structured learning, is used to learn which instantiations of the generalized policy are most likely to lead to a successful grasp in different problem contexts. Experiments are described where Dexter, a dexterous bimanual humanoid, learns to select appropriate grasp strategies for different objects as a function of object eccentricity and orientation. In addition, it is shown that grasp skills learned in this way generalize well to new objects. Results are presented showing that after learning how to grasp a small, representative set of objects, the robot’s performance quantitatively improves for similar objects that it has not experienced before. I.
Evidence-Based Recognition of Objects
- IEEE Transactions on Pattern Analysis and Machine Intelligence
, 1988
"... EL-E: an assistive mobile manipulator that autonomously fetches ..."
Abstract
-
Cited by 6 (0 self)
- Add to MetaCart
(Show Context)
EL-E: an assistive mobile manipulator that autonomously fetches
Humanoid mobile manipulation using controller refinement
- Proc. RSS Workshop: Manipulation for Human Environments
, 2006
"... Abstract — An important class of mobile manipulation problems are “move-to-grasp ” problems where a mobile robot must navigate to and pick up an object. One of the distinguishing features of this class of tasks is its coarse-to-fine structure. Near the beginning of the task, the robot can only sense ..."
Abstract
-
Cited by 5 (1 self)
- Add to MetaCart
(Show Context)
Abstract — An important class of mobile manipulation problems are “move-to-grasp ” problems where a mobile robot must navigate to and pick up an object. One of the distinguishing features of this class of tasks is its coarse-to-fine structure. Near the beginning of the task, the robot can only sense the target object coarsely or indirectly and make gross motion toward the object. However, after the robot has located and approached the object, the robot must finely control its grasping contacts using precise visual and haptic feedback. This paper proposes that move-to-grasp problems are naturally solved by a sequence of controllers that iteratively refines what ultimately becomes the final solution. This paper introduces the notion of a refining sequence of controllers and defines it in terms of controller goal regions and domains of attraction. Refining sequences are shown to be more robust than other types of controller sequences. In addition, a procedure for converting a refining sequence into an equivalent “parallelized ” controller is proposed. Executing this parallelized controller confers all the advantages of iteratively executing the controllers sequentially. The approach is demonstrated in a move-to-grasp task where Robonaut, the NASA/JSC dexterous humanoid, is mounted on a mobile base and navigates to and picks up a geological sample box. I.
Null Space Grasp Control: Theory and Experiments
"... A key problem in robot grasping is that of positioning the manipulator contacts so that an object can be grasped. In unstructured environments, contact positions are typically planned based on range or visual measurements that are used to reconstruct object geometry. However, because it is difficul ..."
Abstract
-
Cited by 2 (0 self)
- Add to MetaCart
(Show Context)
A key problem in robot grasping is that of positioning the manipulator contacts so that an object can be grasped. In unstructured environments, contact positions are typically planned based on range or visual measurements that are used to reconstruct object geometry. However, because it is difficult to measure the complete object geometry precisely in common grasp scenarios, it is useful to employ additional techniques to adjust or refine the grasp using only local information. In particular, grasp control techniques can be used to improve a grasp by adjusting the contact configuration after making initial contact with an object by using measurements of local object geometry at the contacts. This paper proposes three variations on null space grasp control, an approach that combines multiple grasp objectives to improve a grasp. Two of these variations are theoretically demonstrated to converge to force closure configurations for arbitrary convex objects when grasping with two contacts. All variations are found to converge in simulation. Robot grasping experiments are reported that show the approach to be useful in practice.
Grasp Point Optimization by Online Exploration of Unknown Object Surface
- In Proc. Humanoids
, 2012
"... Abstract-In order to realize in-hand manipulation of unknown objects, we introduce an extension to our previously developed manipulation framework, such that long manipulation sequences, involving finger regrasping, become feasible. To this end, we propose a novel feedback controller, which searche ..."
Abstract
-
Cited by 1 (1 self)
- Add to MetaCart
(Show Context)
Abstract-In order to realize in-hand manipulation of unknown objects, we introduce an extension to our previously developed manipulation framework, such that long manipulation sequences, involving finger regrasping, become feasible. To this end, we propose a novel feedback controller, which searches for locally optimal contact points (suitable for regrasping), employing an online exploration process on the unknown object surface. The method autonomously estimates and follows the gradient of a smooth objective function. More concretely, we propose to dynamically switch between manipulability and grasp stability depending on the grasp stability level. Physics-based simulation experiments, involving artificial noise to model real-world sensor readings, prove the feasibility of our approach by rotating an object while readjusting the grasp configuration with all fingers in turn.
1 Real Life Grasping using an Under-actuated Robot Hand – Simulation and Experiments
"... Abstract—We present a system which includes an underactuated anthropomorphic hand and control algorithms for autonomous grasping of everyday objects. The system comprises a control framework for hybrid force/position control in simulation and reality, a grasp simulator, and an under-actuated robot h ..."
Abstract
- Add to MetaCart
(Show Context)
Abstract—We present a system which includes an underactuated anthropomorphic hand and control algorithms for autonomous grasping of everyday objects. The system comprises a control framework for hybrid force/position control in simulation and reality, a grasp simulator, and an under-actuated robot hand equipped with tactile sensors. We start by presenting the robot hand, the simulation environment and the control framework that enable dynamic simulation of an under-actuated robot hand. We continue by presenting simulation results and also discuss and exemplify the use of simulation in relation to autonomous grasping. Finally, we use the very same controller in real world grasping experiments to validate the simulations and to exemplify system capabilities and limitations. Index Terms—Manipulators, Simulation, Robot tactile systems, Modeling
Learning Grasp Strategies Composed of Contact Relative Motions
"... Abstract — Of central importance to grasp synthesis algorithms are the assumptions made about the object to be grasped and the sensory information that is available. Many approaches avoid the issue of sensing entirely by assuming that complete information is available. In contrast, this paper focuse ..."
Abstract
- Add to MetaCart
(Show Context)
Abstract — Of central importance to grasp synthesis algorithms are the assumptions made about the object to be grasped and the sensory information that is available. Many approaches avoid the issue of sensing entirely by assuming that complete information is available. In contrast, this paper focuses on the case where force feedback is the only source of new information and limited prior information is available. Although, in general, visual information is also available, the emphasis on force feedback allows this paper to focus on the partially observable nature of the grasp synthesis problem. In order to investigate this question, this paper introduces a parameterizable space of atomic units of control known as contact relative motions (CRMs). CRMs simultaneously displace contacts on the object surface and gather force feedback information relevant to the object shape and the relative manipulator-object pose. This allows the grasp synthesis problem to be re-cast as an optimal control problem where the goal is to find a strategy for executing CRMs that leads to a grasp in the shortest number of steps. Since local force feedback information usually does not completely determine system state, the control problem is partially observable. This paper expresses the partially observable problem as a k-order Markov Decision Process (MDP) and solves it using Reinforcement Learning. Although this approach can be expected to extend to the grasping of spatial objects, this paper focuses on the case of grasping planar objects in order to explore the ideas. The approach is tested in planar simulation and is demonstrated to work in practice using Robonaut, the NASA-JSC space humanoid. I.
Real Life Grasping using an Under-actuated Robot Hand -- Simulation and Experiments
, 2009
"... We present a system which includes an underactuated anthropomorphic hand and control algorithms for autonomous grasping of everyday objects. The system comprises a control framework for hybrid force/position control in simulation and reality, a grasp simulator, and an under-actuated robot hand equip ..."
Abstract
- Add to MetaCart
We present a system which includes an underactuated anthropomorphic hand and control algorithms for autonomous grasping of everyday objects. The system comprises a control framework for hybrid force/position control in simulation and reality, a grasp simulator, and an under-actuated robot hand equipped with tactile sensors. We start by presenting the robot hand, the simulation environment and the control framework that enable dynamic simulation of an under-actuated robot hand. We continue by presenting simulation results and also discuss and exemplify the use of simulation in relation to autonomous grasping. Finally, we use the very same controller in real world grasping experiments to validate the simulations and to exemplify system capabilities and limitations.
Learning Grasp Strategies Composed of Contact Relative Motions
"... Abstract — Of central importance to grasp synthesis algorithms are the assumptions made about the object to be grasped and the sensory information that is available. Many approaches avoid the issue of sensing entirely by assuming that complete information is available. In contrast, this paper propos ..."
Abstract
- Add to MetaCart
(Show Context)
Abstract — Of central importance to grasp synthesis algorithms are the assumptions made about the object to be grasped and the sensory information that is available. Many approaches avoid the issue of sensing entirely by assuming that complete information is available. In contrast, this paper proposes an approach to grasp synthesis expressed in terms of units of control that simultaneously change the contact configuration and sense information about the object and the relative manipulatorobject pose. These units of control, known as contact relative motions (CRMs), allow the grasp synthesis problem to be recast as an optimal control problem where the goal is to find a strategy for executing CRMs that leads to a grasp in the shortest number of steps. An experiment is described that uses Robonaut, the NASA-JSC space humanoid, to show that CRMs are a viable means of synthesizing grasps. However, because of the limited amount of information that a single CRM can sense, the optimal control problem may be partially observable. This paper proposes expressing the problem as a k-order Markov Decision Process (MDP) and solving it using Reinforcement Learning. This approach is tested in a simulation of a two-contact manipulator that learns to grasp an object. Grasp strategies learned in simulation are tested on the physical Robonaut platform and found to lead to grasp configurations consistently. I.
A Switching Control Approach to Haptic Exploration for Quality Grasps
"... Abstract — Robotic grasping is traditionally approached as a pure planning problem that assumes a priori knowledge of the target object’s geometry. However, we would like our robot to be able to robustly grasp objects with which it has no prior experience. Our approach is to use haptic information t ..."
Abstract
- Add to MetaCart
(Show Context)
Abstract — Robotic grasping is traditionally approached as a pure planning problem that assumes a priori knowledge of the target object’s geometry. However, we would like our robot to be able to robustly grasp objects with which it has no prior experience. Our approach is to use haptic information to drive a search process for appropriate finger contact locations. Given a cost function that is based on the total force and moment applied to the object by the set of contacts, and simple assumptions about the local surface geometry, this search process can be formulated as one of gradient descent on the cost function. Prior work in this area has assumed that the surface of the object local to the contact is either flat or convex. However, when the surface is concave, the search process, in fact, ascends the cost function. Here, we propose a switching controller approach that estimates the local curvature of the object over multiple contacts. This information is then used to switch between one of two methods of estimating the gradient of the cost function. While this new approach shows comparable performance to the original when faced with objects containing only flat or convex surfaces, the new algorithm performs substantially better when objects contain concave surfaces. I.
