Sliding Mode Robust Tracking Control Based on Learning Feedforward Compensation for High Precision Linear Servo System

The permanent magnet linear servo system has some advantages of high speed, high response and direct drive, but the servo performance is influenced by the load disturbances, end-effects, nonlinear friction and the parameters variations. In order to eliminate the influences of the above-mentioned uncertain factors on the basis of ensuring better tracking performance, a robust tracking control strategy combining variable structure control (VSC) with B spline nervous network (BSNN) is presented in this paper. The variable structure control has advantages of the fast response and the invariability of the uncertain factors, but its chattering phenomenon will influence the stationarity of the linear servo system and tracking precision. To reduce the "chattering" input, a BSNN is adopted to eliminate the above-mentioned uncertain factors, thus the steady state accuracy of the system is enhanced further. The simulation results show that the solution not only has stronger robustness to the uncertainty of the linear servo system, but also has better tracking performance