An Exprimental Modelling and Identification of Feed Drive Dynamics with Considering Variable Friction

Abstract

Modeling and accurate identification of the robot dynamics have a direct effect on the implementation quality of control rules. The dynamics of the prismatic joints section, as a part of the robot structure, have a great impact on the overall system dynamics. In order to achieve a high precision dynamic model in the robots, it is important to identify the structure and parameters of the joint’s dynamic model. In this paper, the identification of the feed drive dynamical model as a prismatic joint of a 3-DOF linear independent parallel robot are investigated. Having in mind that friction modeling is an important-nonlinear component of the overall system model, a hybrid method is proposed which is able to identify variable friction parameters. The results of this identification are compared to experimental data obtained from the system. In addition, the results of the proposed method, are compared to the estimation of the friction parameters, which are obtained by the unbiase Least Square(LS) method. From the practical tests it can be inferred that the proposed hybrid method leads to a better output estimate error value.