A new model for nonlinear friction compensation in the force control of robot manipulators

Tracking Control in the Presence of Nonlinear Dynamic Frictional Effects: Robot Extension

“…1, we see that the maximum magnitude of position tracking error for link 1 occurs at large link velocities (i.e., during the transition period of the sinusoidal trajectory). This phenomenon could indicate a possible miscalculation of the viscous friction parameter f d1 in (88) from the procedure outlined in Remark 7; hence, the inaccuracy in f d1 is then relayed to the feed-forward cancellation term F d q of (14) in the exact model knowledge controller of Section IV which may attribute to the relatively minor degradation in tracking performance. In contrast, we see that the maximum magnitude of the tracking error for link 2 coincides with the peaks of the sinusoidally shaped position profiles (i.e., when the link velocity is close to zero).…”

“…where W m (q, q ) was previously defined in (14), and (43) has been utilized. Based on the open-loop structure of (44), we design the torque input τ(t) as follows…”

“…For example in [8], Canudas et al utilized a linear parameterizable friction model to develop an adaptive control algorithm to promote link position tracking for a robot manipulator. Gonzalez et al [14] utilized nonlinear feed-forward compensation for friction while obtaining force control in robotic manipulators. In [2], Armstrong explored the implications of the Stribeck effect on feedback control using a reduced order model of friction.…”