Research On High-precision Positioning Technology Of Robot Terminal Based On Binocular Vision

Robotics,as the fundamental technology of smart manufacturing,has gradually superseded the conventional industrial operating mode with the robust advancement of various countries worldwide.Consequently,robotics has become pivotal technology for boosting production efficiency and minimizing labor expenses.Due to the high repetitive positioning accuracy of the robot end and the low absolute positioning accuracy,the robot end positioning technology is not universal in high-precision assembly,spraying,welding and other fields.This paper focuses on addressing the issue of poor end-point positioning accuracy exhibited by robots,a novel technique utilizing binocular vision is proposed to achieve high-precision positioning of the six-degrees-of-freedom robot end-point.The dissertation chiefly encompasses the subsequent facets:1.Research on eye-to-hand visual positioning technology.In this paper,a binocular vision system and a six-degree-of-freedom robot are used to form an eye-to-hand vision positioning system.The binocular camera is used to locate the corner points of the checkerboard target fixed at the end of the robot,and the robot is indirectly positioned according to the position offset between the feature point and the end of the robot.end position.Experiments have confirmed that the measured camera positioning distance is accurate within 0.05 mm.2.Calibration techniques for robot tool coordinate systems and hand-eye coordination.The offset between the tool target and the end coordinate system of the robot is calibrated,the distance error function is based on relative distances in various coordinate systems,and the optimal solution is estimated using the trust region algorithm.After calibration,the distance accuracy is verified to be 0.1965 mm.Afterwards,the traditional Tsai twostep method and the improved matrix direct product method are used for hand-eye calibration.The experimental comparison verifies that the improved matrix direct product method has a better effect and the position accuracy reaches 0.3268 mm.In this way,the pose conversion relationship between the binocular vision system,the checkerboard grid and the target six-degree-of-freedom robot is established to form a closed-loop visual positioning robot end system.3.Robotic kinematics modelling and parameter identification.In order to address the imprecision of absolute positioning accuracy in industrial robots,a method of identifying kinematic parameters through binocular vision is put forward.The binocular visual system is employed to communicate with the robot and gauge the real distance of various end positions in relation to the midpoint of the sphere,and compare it with the theoretical distance to construct a relative distance regression error function;the enhanced particle swarm optimization technique is applied to evaluate the optimal solution,and the kinematic variables are modified.The experimental findings demonstrate that the average distance deviation is 0.2260 mm,which effectively boosts the absolute positioning accuracy of the robot.Simultaneously,in keeping with the principle of robotic inverse kinematics,the rotation angle of each joint at the point when the robot achieves the intended posture is resolved,thereby facilitating the robot motion.4.After the calibration of each subsystem was completed,a hardware system platform was built for the overall experiment.Using a binocular stereo vision system,the robot’s end-effector was positioned to the desired location based on coordinate system transformation.The actual position of the end-effector before and after calibration was compared with the expected position in the camera coordinate system.Through analysis,it was found that the positioning accuracy of the robot’s end-effector improved from3.7133 mm to 0.9557 mm,representing a 74.26%.increase in accuracy.Figure [46] table [6] reference [79]...