Neural Net-based Modeling and Control of a Micro-positioning Platform Using Piezoelectric Actuators

Abstract

The paper proposes an integrated controller which combines feedforward, PID and a cross-coupling controller to achieve high accuracy for a dual-axis piezoelectric actuated platform. A Preisach model is created to numerically describe the hysteresis behavior of the piezoelectric actuators. On the basis of the Preisach model, a feedforward compensator is developed to compensate for the hysteresis nonlinearity. A PID controller is introduced to further suppress the tracking error due to the modeling inaccuracy and hence to obtain precision tracking control. Finally, a neural net-based cross-coupling control scheme is proposed to reduce the contour errors, which are very typical in dual-axis tracking control problems. The approach developed is numerically and experimentally verified with a variety of operating conditions.