Adaptive neural network finite-time command filtered tracking control of fractional-order permanent magnet synchronous motor with input saturation

Abstract In this paper, the finite-time position tracking control for the fractional-order permanent magnet synchronous motor with input saturation, load disturbance and parameter uncertainties is considered. Based on the command filtered backstepping method, a novel adaptive neural network finite-time controller is presented. First, the neural network is introduced to approximate the uncertain function. Then, by using a command filter at the output side of the virtual signal, the issue of “explosion of complexity” is avoided. In addition, an adaptive law is applied to eliminate the approximation error and filtering error. Meanwhile, by utilizing the terminal sliding mode control technique, the finite-time signal tracking is achieved. Furthermore, an auxiliary design system is constructed to cope with the input saturation constraint. The proposed scheme not only possesses the superiorities of the classical command filtered backstepping, but also holds quick finite-time convergence property. Numerical simulations and experiment results are included to reveal the validity of the proposed design.