Modeling Preisach-type hysteresis nonlinearity using neural networks
对Preisach类的迟滞非线性神经网络建模

In order to approximate the behavior of hysteresis nonlinerity which often severely limits the performance of the system, a neural-network-based hysteresis model is presented in this paper. A novel hysteretic operator is firstly proposed to transform the multi-valued mapping of Preisach-type hysteresis into a one-to-one mapping so that the neural networks are capable of implementing identification for hysteresis. The proposed model has a simple structure and simplifies identification procedure. Moreover, it is convenient to tune the weights of neural networks for the identification of hysteresis in different conditions. Finally the approach is applied to model the hysteresis in piezoelectric actuator and compared with the well-known KP model.