Estimation of grasp type and determining grasping points for Adaptive Grasping Device

The article describes the grasping point algorithm (GPA), aimed to determine points for picking up objects by a manipulator equipped with a proprietary gripper enabling three picking modes (suction cup, two fingers, three fingers). The paper outlines the topic of unknown object picking using various types of grippers, both from a stack of objects and stand-alone items. It describes details of four methods available in the literature. Presented GPA algorithm consists of four separate functions dedicated to appropriate grasping methods chosen based on a decisionmaking process. The dedicated functions has been described in detail and tested. Used methods have been tested manually on 15 different objects. GPA algorithm has been evaluated in realworld tests as well. The presented tests showed efficiency of approximately 92% in the case of moving previously unknown objects from the stack to the target cuvette.