Identification of manipulator kinematics parameters through iterative method

In this study, a numerical procedure for the identification of manipulator kinematics parameters is proposed to determine the precise positioning of a robot. To achieve the objective an iterative-based optimization technique was formulated which minimizes the Cartesian positioning of a manipulator by determining its rigid geometrical parameters such as link lengths, twists, and off-set angles. A closed kinetic chain was formulated using only a internal joint angle reading, which directly provide the calibration mechanism for the manipulator. A gradient projection algorithm was used to obtain the optimal parameters that had satisfied the world coordinates from the angle reading. The proposed method is numerically stable since it converges to the correct answer, with virtually any initial approximation, and also produces the null solution. A case study on the RT100 SCARA type robot was presented