Download

The paper deals with the grasping of partially known objects in an industrial context. The 1 proposed method is based on two steps: 1) the generation of an object model, based on a number of 2 point clouds acquired by means of a depth camera from different point of view; 2) the alignment of 3 the obtained model with the current view of the object in order to detect the grasping pose. More in 4 detail, the model is obtained by merging different point clouds with a registration procedure based 5 on the Iterative Closest Point (ICP) algorithm. The grasping pose is selected on the model. Such a 6 procedure must be executed once, and it is necessary in the presence of objects only partially known 7 or when a model is not available. Then, during the normal operations, the current object view is 8 aligned to the model, by means of the ICP algorithm, and the current grasping pose is estimated. 9 Quantitative experiments using a 7 degrees-of-freedom manipulator on different objects demonstrates 10 the effectiveness of the proposed approach for some kind of workpieces used in the automotive 11 industry.