Fuzzy Logic Scheduling in PI Control of the Servo System Under High Stiction Torque

Abstract

This paper is concerned with a controller design for a servo system with a high stiction torque. A complete friction model is adopted that caters for a dynamic stiction torque. A classical cascaded control structure that is composed of the inner velocity and outer position control loops is employed. The PI controllers are used in this loops. Due to a high stiction torque a high accuracy of servo tracking in the steady-state and a high dynamic performance cannot be simultaneously achieved by fixed parameter controller under uncertainty in the friction model parameters. High gains of the controller needed to reject friction in order to obtain required steady-state error would lead to a loss of stability. Gain scheduling is proposed based on the servo speed measurements. The gain scheduling functions cannot be derived using classical techniques due to the uncertainty and nonlinearities. A fuzzy logic approach is then applied that easily utilises available qualitative representation of the functions. Performance of the controller is illustrated by simulation results.