Design of position controller of linear pulse motor using neural network

In this paper, we introduced the neural network of feedback error learning type to improve position accuracy of a linear pulse motor. A linear pulse motor is direct drive motor that has good performance in terms of accuracy, velocity and acceleration compared with the conventional linear motion system consisting of rotary motor, gear and screw spindles. In general, conventional position controllers of linear pulse motor disregard the modelling error and load variations, which cause inaccuracy in position control. In order to compensate for these factors, we introduced a neural network based on learning feedforward component, which is used to the conventional feedback controller consequently. The simulation results show that the proposed controller works efficiently for accurate position control of a linear pulse motor