Development of strategies to compute force- closure grasps on polyhedral objects

  1. PRADO GARDINI, SIXTO RICARDO
Dirigida por:
  1. Raúl Suárez Feijóo Director/a

Universidad de defensa: Universitat Politècnica de Catalunya (UPC)

Fecha de defensa: 10 de diciembre de 2008

Tribunal:
  1. Luis Basañez Villaluenga Presidente/a
  2. Joan Rosell Gratacos Secretario/a
  3. Pedro José Sanz Valero Vocal
  4. Fernando Torres Medina Vocal
  5. Federico Thomas Arroyo Vocal

Tipo: Tesis

Teseo: 275376 DIALNET

Resumen

The synthesis of force-closure grasp is one of the fundamental problems in the grasp and manipulation of objects by means of mechanical hands. Then, the force-closure analysis and grasp synthesis are fundamental issues in multifingered grasping. However, there are few approaches in the literature dealing with an efficient computation of Force-Closure Grasps (FCG) on polyhedral objects due to complicated geometry and high dimension of the grasp space. In this Dissertation approaches for computing force-closure grasps with three and four frictional contact points on polyhedral objects are proposed. In each approach, first, the sets of faces whose relative orientations and positions allow to obtain force-closure grasps are determined. Second, these sets are evaluated using a quality function and the best one is selected for the grasp; and third, on the selected set faces the fingertip contact points are determined such that they assure the force-closure property. These approaches are based on geometric operation in the 3D physical space and do not present iterative loops. Aditionally it is developed an approach for the generation of force-closure grasps with four frictional contact points that allow that at least three fingers can loose the contact with the object (one finger at a time) without making the resulting three contact points grasp losing the force-closure property. This type of grasp is useful to allow different possibilities for the object manipulation (regrasp without leaving the object). The dissertation includes a comparison of the resulting grasps with the optimum ones in different cases.