Development of a Method for Robotic Sorting of Metal Objects

This paper presents a method for the robotic sorting of objects in automatic mode, which makes use of the Kuka industrial manipulator. The method proposed in this work performs the segmentation of objects using a neural network, the selection and the determination of the spatial coordinates of the picking object, and the movement of the manipulator and control of the picking device. The results of the tests demonstrated that the proposed method was able to accurately determine the spatial coordinates of the object with minimal error. Furthermore, the Euler angles of the object centre were calculated with an error of 1.73° on the X-axis and 1.77° on the Y-axis. The findings indicate the potential for extending the scope of applications for industrial manipulators. The results demonstrate the feasibility of employing these manipulators in intricate and congested environments, enabling the sorting of object.