A Bolt Assembly Robotic System Based on Hybrid Vision: Effect of its Process Parameters on Fastening Performance

Abstract Bolt assembly is very common in modern manufacturing. With the improvement of industrial automation levels, the demand for automatic assembly increases. However, the robotic automation still suffers from various impracticalities. In this work, to improve the environment adaptability and reduce the system cost, a robotic system based on hybrid vision is presented for the bolt assembly, which is composed of a 6-DOF robot, a hybrid vision and an electric fastening tool. In the hybrid vision, a scene monocular vision and a hand-eye monocular vision are combined to obtain the accurate pose of the hole. Besides, motion parameters of the robotic system, including the end-effector linear speed of the manipulator, the rotation speed of the tool, have the major impact on the assembly efficiency and quality. For this problem, the effect of the combined changes in robotic motion parameters on the bolt fastening efficiency and quality was studied based on experimental data. The efficiency and quality were evaluated by the speed of the bolt screwed in and the combination of standard deviation and average value of the torque variation, respectively. Results can provide proper values for motion parameters of the robotic system through considering requirements of the actual bolt assembly.