Study on Application of RBF Neural Network in Control of Giant Magnetostrictive Actuator

摘要

For magnetostrictive actuator (GMA) inherent hysteretic, a new real time hysteresis compensation method consisting of radial basis function neural network (RBF) feedforward and PID feedback controller is presented to achieve the position tracking control of GMA. Simulation results show the control strategy is effective for GMA hysteresis which is changed by the input signal, it can set up the hysteresis inverse model of GMA, thus eliminate the influence of nonlinear hysteresis and achieve high precision control of displacement GMA.