Research On Error Analysis And Compensation Methods Of Serial Robot

With the rapid technological development in the field of robotics,higher and higher requirements have been put forward for the production and use of robots.And it is very important to ensure the motion accuracy of the robot in the operation and use of the robot,so it is necessary to develop a more universal and fast error calibration and compensation method.In order to improve the end positioning accuracy of a robot,the general operating steps are error analysis,error modeling,parameter calibration,error compensation,etc.In order to effectively improve the positioning accuracy of the robot,it is particularly important to establish a complete set of error compensation system.This article focus on the series robot structure parameter calibration and end position compensation problem,analyzed series robot error sources,puts forward a series of robot error modeling method,which is based on the variational method and screw theory,and put forward the concrete algorithm for error calibration and error compensation,also puts forward a position error compensation method which is based on neural network.Specific research contents are as follows:(1)For the 5R series robot,the forward kinematics model of the robot was established by DH method which can extract the parameters of the joint linkage,and algebraic analysis method was used to derive the inverse solution formula of the 5R series robot,and the expression of the joint Angle of the robot was obtained,and the algorithm operation steps of the inverse kinematics were sorted out.(2)Based on the concept of variation,the mapping relationship between the terminal pose error of the series robot and the error source of the robot is constructed.With the help of the spiral theory and the property of the variational space,the relation function between the terminal pose error and the structural error of the robot under the representation of the variational spiral is established.Finally,a general matrix model of pose error of series robots was summarized.Through data simulation,the error was less than 0.2%,which verified the accuracy of the variational error model in the case of small error in error source parameters.(3)The relationship between the robot error parameters and the position error of the terminal coordinate point is studied and analyzed,the error parameter identification model based on the distance error is established,and the algorithm for robot kinematics calibration and positioning error compensation is put forward.Through data simulation,the positionaccuracy of the robot is improved by 92%,which proves the validity of the structure parameter calibration algorithm and the error compensation algorithm.(4)Aiming at the prediction and compensation of robot terminal pose,the construction method of BP neural network is studied and analyzed,and the algorithm of terminal pose prediction and compensation based on neural network is proposed.Set up the training data and test data sets for training the neural network model and test verification,by the neural network toolbox to complete for the establishment of the neural network model and the simulation test,constructed based on neural network algorithm,the end distance error was reduced by 53.1%and 63.5% respectively,show that the algorithm is to predict end position and compensation is effective.