Neural nets based modeling of inverse model for hysteresis using continuous transformation

This paper proposes a novel and simple approach for modeling of the inverse hysteresis. In this method, the continuous transformation technique is utilized to construct an elementary inverse hysteresis model (EIHM), which sets up a one-to-one mapping between the input space and the output space of the inverse hysteresis nonlinearity. Then the output of the EIHM is used as one of the input signals of the neural network (NN) to approximate the inverse behavior of hysteresis. Finally, the proposed method is applied to the modeling of the inverse model of hysteresis inherent in piezoelectric actuator