Robot Offline Programming Technology And Error Analysis For Terahertz Intelligent Detection System

Terahertz wave usually refers to the electromagnetic wave in the frequency range of 0.1～10THz,which has strong penetrability to non-polar composite materials.With the wide application of terahertz waves,terahertz time-domain spectroscopy technology plays an increasingly important role in the field of non-destructive testing.However,since the two-dimensional scanning guide can only move in the X and Y directions,the conventional terahertz time-domain spectroscopy system can only achieve non-destructive testing on flat samples,but it has errors for the detection of curved samples.Therefore,in this study,a terahertz intelligent detection system is set up with the UR10 robot and the terahertz time-domain spectroscopy system as the main body.Calibration and error analysis are performed for each part of the system.The non-destructive testing for curved surface samples made from non-polar composite material is realized by using the offline programming technology of the robot.The specific work includes the following aspects:Firstly,kinematics research of industrial robots is carried out.According to the improved D-H parameter model,UR10 robot is modeled by using Robotics Toolbox.The forward and inverse kinematics equations of UR10 robot are simulated and verified.The trajectory simulation between two points is performed based on joint space and Cartesian space.And the workspace of UR10 robot is simulated based on Monte Carlo method.Theoretical simulation research work lays the foundation for the research of offline programming technology.Secondly,research on calibration and error analysis of terahertz intelligent detection system is carried out.The factors that affect the system errors especially UR10 robot's positioning error are analyzed.The error model is established,the kinematic parameters are calibrated and the errors before and after the calibration about UR10 are compared.Later the calibration experiment on UR10 robot's tool coordinate system is carried out.The digital model of the sample based on inverse technology is reconstructed,and the error between the scanned point cloud and the inverse digital model is studied.Through the calibration and error analysis of the terahertz intelligent detection system,the positioning accuracy of the system is improved.Next,the offline programming technology research of the curved surface sample is carried out.Through compiling Matlab GUI interface by using calibrated kinematic parameters,UR10 robot's model is established to realize functions related to kinematics simulation and trajectory planning.Based on NURBS(Non-Uniform Rational B-Splines)surface normal extraction technology and curved surface detection path planning technology,offline programming module is compiled.Offline programming for the curved surface sample with projection areas of 100 mm × 100 mm and 150 mm × 150 mm is carried out to generate robot executable scripts.Finally,the terahertz intelligent detection system is set up and the non-destructive testing experiment is carried on.The position information of the tool coordinate origin during the detection process is collected and the 3D errors about the sample reconstruction digital model are compared to verify the detection accuracy.The results shows that the average detection error in the 100 mm × 100 m area is 0.8850 mm,and the average detection error in the 150 mm × 150 mm area is 0.8268 mm.The non-destructive testing for the curved sample by the terahertz intelligent detection system is realized.