Fractional Order Integral Sliding Mode Controller Based on Neural Network: Theory and Electro-Hydraulic Benchmark Test

Abstract

Inaccurate model, uncertainties, and valve bias are the main challenges for the controller design of high precision electro-hydraulic servo systems. To achieve a satisfactory tracking control performance under such difficulties, a fractional order integral sliding mode controller based on a radial basis function neural network (RBFNN) is proposed and experimentally tested in this article. In this approach, the RBFNN is used to handle the uncertainties in the model; meanwhile, the disturbance and the inaccurate valve zero point are handled by the robustness of the designed controller. The fractional order integral sliding mode is proposed to deal with the possible fractional order fluid dynamics as well as to improve the tracking precision. The stability of the control system is proved by the Lyapunov stability theory. The experimental results prove the effectiveness of the control method. Quantitative comparisons with a number of experiment control methods show that the proposed method results in a higher tracking accuracy.