Robot kinematic calibration accuracy using genetic algorithm

Abstract

Position and orientation accuracy of the end-effector is affected by the precision of kinematic parameters elements of the robot. Thus, good precision requires good knowledge of robot physical parameters values. However, this condition can be difficult to meet in practice. Hence, calibration techniques can be devised in order to improve the robot accuracy through estimation of those particular parameters. In this paper, the Genetic Algorithm is used to calibrate the robot kinematic accuracy. A kinematic model is formulated and conducted as an optimization problem for serial robot manipulators. The objective is to analyze and evaluate the performance of the GA in optimizing such robot kinematic accuracy. In this algorithm, the errors in the robot parameters represent the parents and offspring population and the error matrix norms represent the cost functions. The convergence and effectiveness of the presented model are demonstrated by a numerical example.