Automatic design algorithm of a robotic end-effector for a set of sheet-metal parts

Robotic end-effectors in automated production lines are specially designed and built for a unique task and for a particular part. An end-effector capable of grasping different parts with different geometries will expand its benefit and reduce costs. This work focuses on the grasp of sheet metal parts for the automotive industry. In order to maximize the use of a single end-effector, this paper proposes a search algorithm for a simple common grasp configuration. Such configuration imply for an end-effector design capable of grasping a set of sheet metal parts. The algorithm maps possible grasps which are candidate to be common. We define a novel quality measure estimating the distribution of the contacts across the sheet metal part. Candidate grasps which have a sufficient quality are mapped into high-dimensional feature vectors. These feature vectors parameterize the geometry of the polyhedron formed by the contacts locations and the direction of the surface normals relative to the polyhedron. Thereby, they describe the design of the end-effector compatible to the grasp. A database is generated for all possible grasps for each part in the feature vector space. A similarity join based on nearest-neighbor search and classification algorithm are used for intersecting all possible feature vectors over all sheet metal parts and finding common ones. Simulations of a 3-finger grasp on four meshed sheet metal parts resulted in a common grasp. Results of the simulations validate the feasibility of the proposed algorithm.