A novel generalized predictive control strategy for AFM piezo-scanners

Abstract

A piezo-scanner is widely utilized as the scanning actuator of an atomic force microscope (AFM) system, which is commonly actuated via an open-loop manner in the X/Y-axis and regulated with a traditional proportional-integral-differential (PID) controller in the Z-axis. Unfortunately, due to various nonlinearities such as cross-coupling and vibration, there usually exists positioning deviation in the Y-axis and stabilization difficulty in the Z-axis, especially when performing high-speed scanning tasks. Moreover, there are some inherent drawbacks of the traditional PID controller in Z-axis, such as low control bandwidth, large overshoot, and so on, which badly limit the scanning/imaging performance of an AFM system. To deal with the aforementioned shortcomings, considering the advantages of the generalized predictive control (GPC) algorithm, two advanced control strategies are developed for the AFM system by utilizing the GPC algorithm for Y/Z-axis, so as to achieve superior tracking performance in fast scanning tasks and then enhance the imaging performance of the studied AFM system. More specifically, in the Y-axis, a combined controller of GPC algorithm and traditional PID algorithm is designed to compensate for the positioning deviation. In the Z-axis, another GPC controller is proposed to replace the traditional PID controller, which sufficiently shortens the transient process. The designed control scheme is fully tested in an AFM testbed, with experimental results clearly verifying the superior performance of the designed strategies.