High-precision positioning of a mechanism with nonlinear friction using a fuzzy logic pulse controller

A new approach to very accurate positioning of mechanical devices with nonlinear (stick-slip) friction is presented in this paper. The proposed controller applies narrow torque pulses to move a mechanism to a desired position despite nonlinear friction. The pulse shapes generated by the controller are computed using a fuzzy logic approximation of the dependence between the desired displacement and the torque pulse shape. The closed-loop stability conditions for the proposed controller are derived taking into consideration a random variation of friction. A detailed experimental study of the system response to different torque pulse shapes and a detailed controller design are presented for a direct-drive mechanism. In the experiments it was demonstrated that the proposed controller achieves positioning accuracy which is within the limits of the position encoder resolution.