Industrial Robots Error Modeling,Parameter Identification And Compensation Based On Lie Algebra

The geometric errors caused by manufacturing and assembly,as well as the joint compliance errors under loads will significantly influence robot accuracy,which limits its applications in the process of heavy duty and high precision.The robot calibration,including the error modeling,identification,measurement and compensation,is an effective way to improve robot accuracy.The robot kinematic model based on Lie algebra is helpful to solve the problem of completeness,continuity and minimality in the error model.For the industrial robot in the heavy-duty and high-precision application,there are still some deficiencies in recent researches on the robot calibration based on Lie algebra.In the geometric error modeling and identification,the joint compliance error makes a big influence on robot calibration.Error models that ignore the influence of joint compliance errors fail to meet requirements of completeness and minimality.The heavy-duty industrial robot with the counterbalancing system(CBS)forms a local parallel mechanism,affecting the joint stiffness modeling and identification.In the error compensation,The present studies on robot calibration based on Lie algebra mainly focuses on the error modeling and identification,and there lacks consistent approach to compensate geometric errors and joint compliance errors.To solve these shortcomings,the main work of this dissertation is as follows:Firstly,an inverse dynamic model of the robot considering the joint compliance is proposed based on Lie algebra.Using the twist representation and exponential operation,a recursive approach for the inverse dynamic model is derived through deducing the high order derivatives of kinematic variables.The inverse dynamic model provides a method for parameter identification of the CBS.With results of the inverse dynamic analysis,the influence of dynamic forces such as Coriolis force and centrifugal force on the joint compliance error prediction is discussed.Subsequently,a methodology for the joint compliance error prediction is proposed based on the inverse dynamic model.The simulation and experiment are performed to validate the accuracy of the proposed method for the inverse dynamic analysis and joint compliance error prediction.Secondly,in order to construct the joint stiffness model and parameter identification of the robot with CBS,this dissertation presents an equivalent joint stiffness model of the CBS.Based on the proposed inverse dynamic model,a new stiffness identification model is proposed by measuring the motor current and position of the end-effector before and after loading.The joint stiffness identification experiment is performed to validate the accuracy of the proposed model and identification method.Thirdly,a unified error model is established by considering geometric errors and joint compliance errors,and the parameter redundancy problem of joint angle error in error modeling is analyzed.The influence of joint angle errors on robot calibration is discussed by the global exponential product formula.A unified error model is proposed based on the separation and re-representation of joint deformation angel errors.In this case,the proposed unified error model could satisfy the basic three requirements of robot calibration,and the parameters identification method is proposed to identify the geometric parameters and joint stiffness.Using the proposed error model,a numerical simulation experiment of parameter identification is performed.Fourthly,combined with results of parameters identification in the unified error model,an approach of error grading compensation is proposed by modifying the joint angle.The mapping relationship between pose errors and joint angle correction is established by means of the differential property of the exponential product,and an iterative compensation method for geometric errors is constructed.For the joint compliance error,based on the prediction model of joint compliance errors,a methodology is presented to compensate joint compliance errors.The simulation experiment of error grading compensation is carried out to validate the accuracy of the proposed method.Finally,a calibration experiment for the 6-degree-freedom industrial robot HH-150-2 is performed to validate the proposed method of unified error model,parameter identification and graded compensation.A software for unified error modeling,parameter identification,error evaluation and graded compensation is developed using the Matlab/GUI programming environment.