Disturbance observer-based model prediction control with real-time modified reference for a piezo-actuated nanopositioning stage

Abstract

Piezo-actuated micro-/nanopositioning systems have been widely employed in diverse high-precision positioning applications. However, the inherent hysteresis nonlinearity seriously deteriorates the tracking performance of piezo-actuated stages. This paper presents the design, analysis, and validation of a novel control scheme termed model prediction control (MPC) with real-time modified reference based on disturbance observer (DOB) to suppress the hysteresis nonlinearity and model uncertainty, in which the nonlinear effects are treated as an unknown disturbance to the system. In order to remove the most of the interference and diminish the effect of noise, a DOB is designed for the non-minimum phase (NMP) system. Then the difference between the actual displacement and the output of the nominal model termed the residual error is estimated and used to modify the reference in real time for a better performance. By the proposed method, the model of the inherent hysteresis is not required and the controller is established based on the identified nominal model. Its effectiveness is validated through experimental investigations on a commercial nanopositioner. Experimental results show that the proposed method can improve the tracking performance of the piezo-actuated stage, as compared with the traditional MPC and DOB-based MPC.