The Analysis Of Positioning Accuracy And Research On Error Distribution Of Scara Industrial Robot

This paper provides a research on the position accuracy and error distribution of SCARA industrial robot. For position accuracy, all the factors that contributing to the end-effctor error are transformed into the link parameters error to make an analysis. A coordinate for every link of SCARA robot is set up by D-H method. Because of the parrellel of each two axises of link joints, five parameter position and orientation error model is built for SCARA robot. Five Jacobian matrixes,which connecting link parameter error and robot’s end effector error, are built by differential transform method. The above-mentioned five parameter position and orientation error model is built based on the five Jacobian matrix. Because error of link parameters come from an assigned range, the error could be uniformly selected from the range. Then apply the selected error into the five parameters error model. Bigger components of end-effector position and orientation error vector can be derived from the combination of orthogonal testing method and the five parameters error model. A quantity of working points are selected evenly from the working space of SCARA robot. Every working point corresponds to an end-effector position and orientation error vector. Choose a proper error vector for every working point by the above method. Apply principal component analysis into the chosen vectors, and two principal components are obtained. Based on contribution rate and coefficients of two principal components, significant effect of link parameter error on robot’s end-effector position and orientation error can be got. So when improving the accuracy of a robot, remarkable affecting factors that contributing to the end-effctor error should be given priority.For error distribution, Genetic Algorithm is used to model an optimization model. Width of fourteen link parameters tolerance zone is obtained respectively by the optimization model. Manufacturing cost of fourteen link parameters tolerance defines the optimization model’s objective function, while end-effector error probability and width of fourteen link parameters tolerance zone are within the bounds defines the optimization model’s constraints.Based on position accuracy analysis, error distribution, and principal component analysis, a comprehensive score is used to describle error distribution of SCARA robot’s working space. Compared with the inside of working space, position and orientation error from the outside is smaller based on several error distribution sections. So the robot should avoid working in the marginal area of the outside.