Research On Kinematic Calibration Of Industrial Robots Based On Vision

In recent years,industrial robots have become an indispensable part of all walks of life,including high-precision industries such as the aerospace industry,and high-precision processing has put forward high positioning accuracy requirements for industrial robots,and improving absolute positioning accuracy has become the main research problem.Under normal circumstances,the widely used kinematic calibration method is to first measure the data with a laser equipment,and then use the least squares method to complete the parameter identification,which has high measurement accuracy,but the cost is very expensive and the calibration process is difficult.Therefore,this thesis adopts vision-based measurement,combined with the distance error model,to avoid the error caused by the coordinate system conversion,and the artificial bee colony algorithm is selected for parameter identification,which improves the positioning accuracy of the robot,and the main work content is as follows:The kinematics of the robot are modeled,and the robot end error model is derived from the differential motion between the connecting rods.The distance error model is established,and the relationship between the distance error and the positioning error is derived.In this thesis,IRB120 is used as the experimental object to verify the kinematic model.The measuring device uses a monocular camera,and the measurement method is the eye on the hand.The calibration plate was selected by using the alumina checkerboard,the corner point extraction and reprojection error analysis were carried out,and the camera’s internal parameters and measurement errors were calculated,and the results showed that the accuracy of the camera’s internal parameters and measurements met the experimental requirements.Aiming at the problem of low accuracy of traditional hand-eye relationship matrix methods,this thesis proposes an improved algorithm based on matrix direct product.The redundancy parameters of the robot were analyzed,20 parameters that needed to be identified were obtained by the QR decomposition method,and the number of conditions and observability index of the error matrix were analyzed to clarify the number of robot terminal poses that need to be collected in the experiment.In order to reduce the interference of noise,the observability index O2 was selected to determine the optimal set of measurement pose points,and the least squares method was used for parameter identification.In view of the shortcomings of the least squares method,the artificial bee colony algorithm is applied to parameter identification to further improve the positioning accuracy of the robot.The results show that the positioning accuracy of this method is improved compared with the least squares method.The software system of the robot calibration system platform is designed,including communication components,visual components,motion control components,etc.According to the results obtained by parameter identification,the geometric parameters of the robot are corrected,and the accuracy and reliability of the identification results are further verified by randomly taking points and calculating the distance error accuracy.