Neural networks Preisach model and inverse compensation for hysteresis of piezoceramic actuator

The hysteresis nonlinear characteristic of the nanometer positioning system based on piezoceramic actuator decreases the accuracy of the nanometer positioning stage seriously. To compensate the hysteresis nonlinearity and improve the precision of system with hysteresis, this paper studies the modeling of hysteresis and the corresponding inverse compensation. First, a new sorting & taxis model of hysteresis is realized using neural network to describe the hysteresis of the piezoceramic actuator. A BP neural network is introduced to solve the Function F. With this method the error resulting from interpolation is avoided. Secondly, another neural network is promoted do describe the inverse model of hysteresis. The neural network is used in inverse-modeling to replace the reverse checking and interpolation in traditional method, and the hysteresis modeling error is reduced. At last, the inverse Preisach model based on neural networks in used to compensate the hysteresis nonlinearity. Through the experimental results, the effectiveness of the neural networks hysteresis model and inverse model for the piezoceramic actuator is demonstrated. Also the nonlinear characteristic is reduced effectively by the inverse compensation with neural networks.