Comparison study of hysteresis compensation of piezostage using feedforward combined with feedback control

Abstract

Hysteresis is an undesired phenomenon which degrades the positioning precision of a piezostage. Therefore, the hysteresis suppression is required for accurate micro/nanopositioning applications. This paper conducts a comparison study of three different approaches for the hysteresis modeling and suppression. Bouc-Wen model is a typical approach used to describe hysteresis. As a revised version, asymmetric Bouc-Wen model is employed to combat some of its disadvantages. Relevance vector machine (RVM) hysteresis model is proposed to capture the hysteretic behavior by formulating the hysteresis modeling as an on-line nonlinear regression problem. Based on the three hysteresis models, a feedforward control is formulated to compensate for the hysteresis effects. Moreover, a sliding mode control (SMC) is applied as the feedback controller to regulate the remaining errors. The feasibility and effectiveness of the proposed approaches are verified through experimental studies.