Adaptive novel MSGA-RBF neurocontrol for piezo-ceramic actuator suffering rate-dependent hysteresis

摘要

Piezo-ceramic actuators are widely used in micro-electro-mechanical systems. An adaptive inverse neurocontrol design is proposed for positioning control of the piezo-ceramic actuator (PA). Piezo-ceramic actuators exhibit rate-dependent hysteresis which changes its hysteretic behavior when the rate or the frequency of its driving signal varies. Radial basis function neural network (RBFNN) is utilized to model the input/output relation of the PA with rate-dependent hysteresis to make it work as an accurate hysteretic model for positioning control. And an adaptive inverse nonlinear controller is introduced for the rate-dependent hysteresis compensation of PA. A novel membrane structure genetic algorithm (MSGA) is proposed for the adaptive inverse nonlinear neurocontrol design of PA. Compared with the classical genetic algorithm, the experimental results illustrate the performance of the proposed control system as well as the efficiency of adaptive membrane structure genetic algorithm for positioning control of PA system. (C) 2019 Elsevier B.V. All rights reserved.