A Deep Serial Model and Predictive Control for Piezo-actuated Positioning Stages

Fei Dong,Hongyang Xie,Qinglei Hu,Keyou You,Jianpeng Zhong

DOI: https://doi.org/10.1109/tmech.2024.3454514

2024-01-01

Abstract:The inherent hysteresis nonlinearity of piezo-actuated positioning stages (piezo-stages) is very difficult to deal with due to the amplitude- and frequency-dependent characteristics, which severely limits the tracking accuracy for high speed trajectories. In this article, we first develop a deep serial model to describe the dynamics of the piezo-stage by using historical voltage-displacement data over a period of time. It achieves relative prediction errors less than 0.16% on sinusoidal trajectories with frequencies greater than 72% of the resonance frequency of the piezo-stage through an elaborately designed network structure that includes direct connections between the input layer and the output layer. Then, we design an integral model predictive control (iMPC) and build a feedforward neural network (FNN) to learn its optimal solution offline. This forms the proposed FNN-iMPC and ensures the feasibility of evaluating the control law within the sampling time of 0.1 ms. It achieves a maximum positioning error of 0.02 mu m for a +/- 32 mu m staircase reference signal and a maximum tracking error of 0.19 mu m for a sinusoidal reference signal with a range of 10 mu m and a frequency of 500 Hz in real experiments.