GMS friction model identification and compensation based on velocity observer

Friction is a kind of strong nonlinear factor that severely degrades accuracy of servo system. The fiction compensation based on model can effectively predict the friction amplitude and compensate the dynamic error. The friction is predicted by the generalized Maxwell-slip model (GMS). In order to improve the parameters identification accuracy of friction model, and overcome the error effects of the low speed velocity measurement, the full-order speed observer was designed as feedback speed information. Based on this observer, an experimental parameter identification method of GMS friction model was implemented. In order to verify the proposed identification method effectiveness for friction compensation, some dragging experiments on a late-model space large end-effector were done. Results of experiments show that the precision of the position tracking can reach 0.02 mm by this friction compensation, increased more than 30% compared with Stribeck friction model which has fixed parameters.