Design Of The Calibration System Based On Draw-wire Sensors For Industrial Robot

Nowadays industrial robots have been widely used in many complex industrialapplications, the demand for the need of high accuracy robot is more and more great.And we can improve the accuracy through calibration with no structure changed.Based on draw-wire sensors, we have developed a new kind of industrial robotcalibration system to calibrate the zero values of joints, the geometric parametererrors and the mechanical transfer errors.To calibrate the OTC6-axes industrial robot, we have built four kinds ofkinematic models, the D-H model, the modified D-H model, the simplified D-Hmodel and the D-H model with joint variable ratio k, based which we built thedistance error models to identify and compensate the model parameter errors. Andwe developed the calibration software in matlab to identify the parameters. At last,we compensated the parameters to the corresponding kinematic model. We find thatthe joint variable ratio k has an great effect on improving the accuracy of theend-effector.Comparaed with the current robot measurement products, we proposed theperformance index of the new calibration system. To measure the whole workspaceof the robot and the economy, we applied the draw-wire sensors in this system. Thedraw-wire sensor is designed to measure liner displacement. So we designed a kindof steering mechanism and the adaptor between steel wire and the end-effector tomeasure the position of the end-effector more freely and flexibly. And by the use ofX4encode model of the PCI-1784U data acquisition card,which is producted byADVANTECH, we further improved the system resolution. We bulit themathematical model,and according to the unique structure of the system,wedesigned a simple method to solve the nonliner equations. It was proved to improvethe efficiency of solving the problem by four iterations.Based on the mathmatical model, we set up the system error model andanalysed the effect of structure parameters on the measurement accuracy, by whichwe decided the structure parameters and drew the workspace of the calibrationsystem. Then we analysed the measurement error from five aspects, such as theradius error, the rotation accuracy of axis, the measurement error of draw-wiresensors, the deformation of mechanism and the axial misalignments. The totaldesign accuracy can be as large as0.14mm, which is meet the requirement of thesystem performance.After the machanism is assembled and the hardware connected, the calibrationsystem is designed in labview. When the whole system is set up, we have calibrated the structure parameters and the pulse equivalent. We have tested thestatic accuracyof the system, which is not as well as whtat we needed. There are many structureproblems we need to improve to realize the accuracy that we have designed.