Research on neural networks based modelling and control of electrohydraulic system

The electrohydraulic servo system of a certain type of mines weeping plough is a complex and nonlinear system. It is difficult to construct its accurate model by first principle method and to achieve satisfactory control performance by traditional PID controller. In this paper, the radial basis function neural network with orthogonal least square learning algorithm is used to model the electrohydraulic system and the neural network based direct inverse is adopted to control the system. The experimental results and comparisons with other techniques clearly show the validity of the proposed methods.