Dynamic Parameter Identification of a 6-DOF Industrial Manipulator Considering Friction Model

Abstract

Dynamic parameters identification is essential for motion control of industrial robots. Precise identification is still an open issue due to friction, transmission backlash, and measurement noises of sensors. In this paper, based on a minimum set of parameters considering friction model, a unified identification approach is proposed. With this method, some essential data from robot's joints motion can be sampled and calculated at the same time, thus all dynamic parameters are also acquired at one time. The 5-th order Fourier series is used as the excitation trajectory. The optimal excitation trajectory is determined by minimizing the condition number, and the minimum set of parameters of the robot is obtained by combining with the least square method. Experiment results taken on a 6-DOF industrial manipulator Staubli TX-60 verified the correctness of the presented method.