Modeling and SIMO controlling of piezoceramic hysteresis

Micro-displacement devices, especially nano-scale actuators based on the inverse piezoelectric effect of piezoelectric ceramic are widely used. In Atomic Force Microscope (AFM) nano-level lateral resolution of probe or sample micro-displacement can be achieved using piezoelectric actuator stage. However, significant accuracy reduction is brought about by nonlinearity and multiple-value characteristics of piezoceramic hysteresis. In order to enhance the resolution of AFM system, the modeling of piezoelectric hysteresis using BP neural-network is presented in this paper based on the central symmetry characteristics, and the model parameters are gained by means of neural network training, then a Single-Input-Multiple-Output (SIMO) control method of piezoelectric ceramic is constructed. Based on the SIMO control model the open-loop tracking control experiment for piezoelectric ceramic is performed, and the tracking control error is between -47nm and 63nm. The experiment results show that the control model has the advantages of high open-loop tracking accuracy and anti-interference capability.