Adaptive Nonsingular Terminal Sliding Mode Control for Electro-Hydraulic Servo Systems Based on RBF Neural Network and Nonlinear Tracking Differentiator

To address the challenges of non-smooth reference force commands, uncertain nonlinear dynamics, and external load disturbances in electro-hydraulic servo systems (EHSSs), this study proposes an adaptive nonsingular terminal sliding mode control (ANTSMC) method combined with a nonlinear tracking differentiator (NTD) and a radial basis function neural network (RBFNN) to improve the tracking performance of the loading force. First, the NTD is adopted to generate smooth reference force signals and their corresponding derivatives. Subsequently, an RBFNN is developed to approximate the unknown lumped uncertainty online, and its estimate is incorporated into the control law for adaptive uncertainty compensation. On this basis, an ANTSMC law is designed to improve force tracking performance and attenuate residual approximation errors. Furthermore, the closed-loop boundedness and convergence properties are investigated through Lyapunov stability analysis. Finally, simulation and experimental results demonstrate the effectiveness of the proposed method.