Posture Error Correction of a Six-DOF Serial Manipulator Based on Genetic Algorithms

Traditionally, the posture errors of a serial manipulator are corrected through the calibration of geometric parameters. However, this paper presents a compensation method based on genetic algorithms which corrects the joint variables of a serial manipulator, thus reducing the posture errors. The structures of the kinematics model and the error model of the 6-DOF manipulator are firstly established. Using the position and posture data of the manipulator measured by the NDI 3D dynamic displacement measurement system, the errors between the desired and the measured poses of the end-effectors are then calculated and set as the fitness function. The fitness function is used in the genetic algorithms to determine the compensation for the manipulator’s rotational joint variables. Finally, the joint variables ‘compensation is implemented to the control system of the designed 6-DOF serial manipulator, and groups of experimental data are presented to demonstrate the efficiency of the technique discussed and the improved performance of the manipulator.