Neuro-intelligent networks for Bouc-Wen hysteresis model for piezostage actuator

Piezoelectric stage has become promising actuator for wide applications of micro-/nano-positioning systems represented mathematically with Bouc-Wen hysteresis model to examine the efficiency. In this investigation, the numerical study of piezostage actuator based on nonlinear Bouc-Wen hysteresis model is presented by neurocomputing intelligence via Levenberg-Marquardt backpropagated neural networks (LMB-NNs). Numerical computing strength of Adams method is implemented to generate a dataset of LMB-NNs for training, testing and validation process based on different scenarios of input voltage signals to piezostage actuator model. The performance of LMB-NNs of nano-positioning system model is validated through accuracy measures on means square error, histogram illustrations and regression analysis.

Automated summary

Piezoelectric stages are promising actuators for micro-/nano-positioning systems, represented mathematically with the Bouc-Wen hysteresis model. This study utilizes neurocomputing intelligence and Levenberg-Marquardt backpropagated neural networks to numerically investigate a piezostage actuator based on the nonlinear Bouc-Wen hysteresis model, using the Adams method to generate a dataset for training, testing, and validation of the LMB-NNs. The performance of the nano-positioning system model is validated through accuracy measures, histogram illustrations, and regression analysis.