Parameter Identification of a Friction Model for Robotic Joints

Abstract

In this paper a new identification method to obtain the friction parameters in the joints of robotic manipulators is presented. These parameters are coulomb friction, static friction, Stribeck velocity constant and viscous damping coefficient. The available methods to find these parameters either require the design of a controller or the precise value of system parameters such as mass moment of inertia. In contrast, the new method proposed here finds these parameters by a nonlinear optimization approach which requires neither any knowledge of system’s parameters nor any controller design. The corresponding nonlinear optimization problem is solved using an efficient technique which does not require iteration or any initial estimate of optimization parameters. The new method proposed in this paper was experimentally verified on a robotic manipulator available in the robotics laboratory at the University of Saskatchewan.