Data Driven Adaptive Control with Hysteresis Input for a Piezo-Actuated Stage

As a core component in the micro/nano-manufacturing, the piezo-actuated stage has superior performance of ultra-precision positioning, fast response and compact size. However, the inherent nonlinearity, such as hysteresis and creep characteristics, in the piezo-actuated stage impacts its positioning precision severely. To eliminate the impact of the nonlinearity of a piezo-actuated stage, a data driven adaptive control method with hysteresis input is proposed in this paper. An unknown nonlinear function with hysteresis input is adopted to depict the piezo-actuated stage. Then, the undetermined parameters of the controller are adjusted online by using the input and output measured data. Finally, the availability of the proposed control method have been demonstrated by a series of experiments compared with a hysteresis inverse compensation-based model reference adaptive control method and a classical model free adaptive control method.

1 Introduction

With the evolution of ultra precision-manufacturing, piezo-actuated stages have become widespread, because of their high resolution, good rigidity and rapid response [1], [2]. However, due to the inherent nonlinearity of piezo-actuated stages, the positioning precision is decreased ineluctably. Hysteresis is the representative characteristic for piezo-actuated stages. As shown in the Fig. 1, the current output of piezo-actuated stages not only dependents on the current input but also relay on the previous input and the previous output [3]–[5]. Moreover, the asymmetry and rate-dependent nonlinearity exist in the input-output loop for piezo-actuated stages [6]. Control design for mitigating the effects of these characteristics is a challenge that has received much attention lately.