A online-trained neural network controller for electro-hydraulic servo system

Abstract

A multilayer neural network supervised online controller based on the Levenberg-Marquardt training algorithm is proposed for the tracking control problem of the electrohydraulic position servo systems subjected to constant and time-varying external load disturbances. The Levenberg-Marquardt algorithm is a combination of the steepest decent algorithm and Gauss-Newton algorithm. Compared with the conjugate gradient algorithm and variable learning rate algorithm, the Levenberg-Marquardt algorithm is much more efficient than either of them on the training steps and accuracy. The control strategy is used to adapt uncertainties of disturbances and learn their inherent nonlinearities. Simulation results illustrate that the neurocontroller used in supervised control schemes can result in good robustness and tracking property.