Particle Swarm Optimization for Identification of GMS Friction Model

This paper addresses the identification problem of the generalized Maxwell-slip (GMS) friction model. The GMS model is a dynamic friction representation capable of describing essential friction characteristics. However, the identification process is complicated by the presence of nonlinearly-occurring parameters, the hybrid structure of the GMS model, and lack of accurate friction force measurements. Therefore, an adaptive friction compensator is developed, based upon a linearly-parameterized version of the GMS model, that provides estimates of friction forces for trajectory tracking and identification purposes. The Particle Swarm Optimization (PSO) method is then employed to identify the nonlinear GMS model using these friction force estimates. Numerical simulations are performed to illustrate the validity of the proposed approach