Complementary Sliding Mode Control via Elman Neural Network for Permanent Magnet Linear Servo System

Abstract

The permanent magnet linear servo system is usually susceptible to uncertainties, such as parameter variations, external disturbances, and friction forces. To address this problem, a complementary sliding mode control (CSMC) via Elman neural network (ENN) was presented in this paper. First, the mathematical model of the permanent magnet linear synchronous motor (PMLSM) with a lumped uncertainty was established. Second, on the basis of the traditional sliding mode control (SMC), CSMC was designed by combining the integral sliding surface with the complementary sliding surface. CSMC is generally used to reduce the chattering phenomenon and, consequently, to improve the tracking performance. However, the values of the switching gain and the boundary layer thickness are difficult to select in CSMC. To deal with this problem, ENN was adopted in the proposed CSMC system to replace the switching control law. Due to its strong learning ability, ENN can estimate the value of the lumped uncertainty and adjust the parameters online, thus further improving the robustness of the system. In addition, to verify the control performance of the proposed method, a digital signal processor (DSP) was implemented as the experimental platform to control the mover of the PMLSM for the tracking of different reference trajectories. The experimental results show that the proposed control strategy not only improves tracking accuracy but also guarantees the robustness of the system.