Bard, C., Laugier, C., Milesi-Bellier, C., Troccaz, J., Triggs, B. and Vercelli, G.: 1995, Achieving dextrous grasping by integrating planning and visionbased sensing, International Journal of Robotics Research 14(5), 445-464.  Home/Search   Document Details and Download   Summary   Related Articles   Check

.... three dimensional approach, although divergent stereo has also been used (Santos Victor, Sandini, Curotto and Garibaldi 1995) Other schemes such as a camera free to rotate about the scene independently of the robot end e ector have also been used (Kinoshita 1998) or mounted on a second robot arm (Bard, Laugier, Mil esi Bellier, Troccaz, Triggs and Vercelli 1995). This work uses the principle of teaching by showing, in which the supervisor shows the system the correct location by placing the robot in the desired relative pose to the target. The system then learns this pose by storing sucient reference information, captured whilst in that pose, to ....

Bard, C., Laugier, C., Milesi-Bellier, C., Troccaz, J., Triggs, B. and Vercelli, G.: 1995, Achieving dextrous grasping by integrating planning and visionbased sensing, International Journal of Robotics Research 14(5), 445-464.

....polyhedral and simple curved objects, using data from two CCD cameras and haptic exploration to give 3D edges and 2D regions. Lyons [6] 7] given the desired distances between digits, used potential fields to shape the hand, but does not derive the desired distances from actual data. Bard et al. [1] used preshaping to plan power grasps on objects. The objects are modeled using elliptical cylinders, which are especially suitable for planning power grasps it is one of the few systems which provides a vision system tailored to grasp planning. The preshapes are planned using a set of ....

C. Bard, C. Bellier, C. Laugier, J. Troccaz, G. Vercelli, and B. Triggs. Achieving dextrous grasping by integrating planning and vision based sensing. Technical report, LIFIAIRIMAG, 1993.

....giving a cell dimension of around 4:7cm. All experiments in this section were done on an SGI Iris R4400. The computation of the occupancy volume of the new object is 2 A set of points within or on the boundary of the discretized cell computed using software from the Esprit SECOND project [1]. It is envisaged, as outlined in the future work below, to connect that part of the system to a viewpoint planner such as described in [8] The vision data, in the form of a voxmap, is then input and merged with the discretized model of the workspace. The blocked paths are determined and are ....

C. Bard, C. Bellier, J. Troccaz, C. Laugier, B. Triggs, and G. Vercelli. Achieving dextrous grasping by integrating planning and vision based sensing. Int. J. Robotics Res. Accepted, to appear in 1995.

....and simple curved objects, using data from two CCD cameras and haptic exploration to give 3D edges and 2D regions. Lyons [6, 7] given the desired distances between fingertips, used potential fields to shape the hand, but does not derive the desired distances from actual data. Bard et al. [1] used preshape to plan power grasps on objects. The objects are modelled using elliptical cylinders, which are especially suitable for planning power grasps. The preshapes are planned using a set of heuristics based on the properties of the elliptical cylinders. When using the precision ....

C. Bard, C. Bellier, C. Laugier, J. Troccaz, G. Vercelli, and B. Triggs. Achieving dextrous grasping by integrating planning and vision based sensing. Technical report, LIFIA-IRIMAG, 1993.

....well adapted to the problem and as simple as possible. For geometric motion planning CAD style geometric representations are optimal and we have developed techniques to extract these from volumetric maps obtained from stereo data, with the emphasis on model simplicity rather than exact fidelity [1]. For grasp planning we have developed an original hierarchical representation based on elliptical cylinders that places a strong emphasis on graspability (i.e. morphology for gripping and accessibility) 2. Preshape selection and grasp determination. To make grasp planning with many degrees ....

C. Bard, C.Laugier, C. Mil'esi-Bellier, J. Troccaz, B. Triggs, and G. Vercelli. Achieving dextrous grasping by integrating planning and vision based sensing. In Int. Jour. of Robotics Research, 1994. To appear.

....classical grippers (see [14] for an overview of this work) Additional work has also been done for extending the capabilities of the previous approaches, in order to be able to deal with dextrous hands. The related solutions are generally based upon the concept of hand preshaping (see [14] and [3]) But, all these geometric based approaches do not completely solve the problem because the stability conditions are partially evaluated using some heuristics and very simple geometric criteria. Some other contributions have focused on the stability problem. In this case, the solutions which have ....

....cylindrical, are also available. The second step executes the grasping using a physical planner . It uses a physical motion generator to reach the object O with the hand H and to ensure the stability of the set H=O. The preshape planner. In our precedent work about grasp planning [4] 5][3], we have proposed to solve the grasp planning until the closure step using the concept of preshaping (concept come from the study of the human grasping) The system is based on the search of geometrical homogeneous areas and their morphological properties usable for grasping. The preshape planner ....

C. Bard, C. Bellier, J.Troccaz, C. Laugier, B. Triggs, and G. Vercelli. Achieving dextrous grasping by integrating planning and vision based sensing. In International Journal of Robotics Research, 1994. Accepted. To appear in 1994.

....are also provided by this step. The second step executes the grasping using a physical planner . It uses a physical motion generator to reach the object O with the hand H and to ensure the dynamic equilibrium of the set H=O. The preshape planner. In our previous work about grasping [3][2], we have show how to solve the grasp planning problem before considering the closure step. The approach is based on the concept of preshaping (concept coming from the study of the human grasping) The basic idea consists in searching for geometrical homogeneous areas and their morphological ....

C. Bard, C. Bellier, J.Troccaz, C. Laugier, B. Triggs, and G. Vercelli. Achieving dextrous grasping by integrating planning and vision based sensing. In International Journal of Robotics Research, 1994. Accepted. To appear in 1994.

....two robot workcell. The planner is optimized for interventionstyle robotics in potentially cluttered workspaces and can be configured for a variety of tasks. For example, it has been integrated into a visually guided grasping system being developed under the European Esprit collaboration SECOND [4], where 0 To appear in IEEE International Conference on Robotics Automation, Nagoya, Japan, 1995. it is used to choose viewpoints for workpiece verification and pose correction and for visual servoing of the grasp approach. Viewing positions are chosen by minimizing a heuristic viewpoint ....

....the range of function values that might occur in a region, that can be used to concentrate search effort on relatively promising regions. 6 Implementation Experiments Our planner is written in C and runs under the ACT robot modelling system [8] as part of our integrated workcell demonstration [4]. The results of a typical run of the system are illustrated in fig. 1. The task was simply to get a clear view of the central peg in the right hand box. The search was over a large 3D space of upright camera placements directed towards the peg and took about 10 seconds on a Silicon Graphics R4000 ....

C. Bard, C. Bellier, J. Troccaz, C. Laugier, B. Triggs, and G. Vercelli. Achieving dextrous grasping by integrating planning and vision based sensing. Int. J. Robotics Research. Accepted, to appear in 1995.

....work has also been done for extending the Motion Planning Combining Geometric and Physical Models. capabilities of the previous approaches, in order to be able to deal with dextrous hands. The related solutions are generally based upon the concept of hand preshaping (see [28] and [2]) But, all these geometric based approaches do not completely solve the problem because the stability conditions are partially evaluated using some heuristics and very simple geometric criteria. Some other contributions have focused on the stability problem. In this case, the solutions which ....

.... stable contact relations between the fingers and the object to be grasped (in this case, q goal is expressed in terms of contact relations involving features of the object and of the hand) Processing such constraints can be done using geometric approaches (see for instance [25] 4] and [2]) The constraints CO d to consider when generating executable motions are the non holonomic kinematic constraints, the dynamic constraints, and the constraints imposed by the forces and the torques which are created by both the vehicle terrain interactions and the control strategy to apply. In ....

[Article contains additional citation context not shown here]

C. Bard , C. Laugier , C. Milesi-Bellier, J.Troccaz, B. Triggs , G. Vercelli: "Achieving dextrous grasping by integrating planning and vision based sensing ", International Journal of Robotics Research. To appear in 1994.

....being a significantly extended rewrite of the single viewpoint planner described in [21] It can be configured for a variety of tasks. For example the first version of the planner was integrated into a visually guided grasping system developed under the European Esprit collaboration SECOND [4, 11], where it was used to choose viewpoints for workpiece verification and pose correction and for visual servoing of the grasp approach. The central activity of the system is the selection of viewing positions and the associated camera robot poses on the basis of task, camera, robot and ....

C. Bard, C. Bellier, J. Troccaz, C. Laugier, B. Triggs, and G. Vercelli. Achieving dextrous grasping by integrating planning and vision based sensing. Int. J. Robotics Research. Accepted, to appear in 1995.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC