The sensitivities of industrial robot manipulators to errors of motion models' parameters

Abstract

In this paper it is focused on sensitivity analyses of position and orientation coordinates of manipulator gripper to errors of kinematics' parameters. Also it is focused on sensitivity analyses of mass forces to errors of dynamics' parameters. In these models the transform graphs are used. These graphs were presented by the author of this paper in The First Int. Conf. Graphs & Mechanics. Practical methods of kinematics' calibration make use of the linear differential error of the kinematics' model. This model is based on the Jacobian of the direct kinematics' model with respect to parameters of this model. These methods use the theory of minimum-square identification of kinematics' parameters too. Practical methods of dynamics' calibration make use of the dynamics models, which are in a linear form in relation to dynamics' parameters. These methods use the theory of minimum-square identification of dynamics' parameters. Measurements are one of the steps of calibration. These measurements should be carried out in the manipulator's states in which the errors of kinematics' parameters cause the largest errors of the gripper's positioning or errors of dynamics' parameters cause the largest errors of the mass forces. These states are named as the most sensitive to errors of kinematics' parameters or as the most sensitive to errors of dynamics' parameters. In this paper measures of sensitivity of the gripper's positioning to errors of kinematics' parameters, and measures of sensitivity of the mass forces to errors of dynamics' parameters are proposed. The example for the calculation of sensitivity of the gripper's positioning for IRb-6 manipulator is presented too.