Active fault tolerant control for high-precision positioning of a non-contact mode uncertain atomic force microscopy

Abstract

A non-contact mode atomic force microscope with chaotic dynamics may exposed to unknown faults, disturbances or uncertain parameters that are not always be compensated using classical control methods. Therefore, a fault tolerant controller must be designed for accurate tracking of the tip-position of the end-effector. In this paper, first, an unscented Kalman filter is designed for joint estimation of the states and parameters for an atomic force microscopy under process noise. The velocity of the end-effector, sample height and unknown fault are simultaneously estimated by measuring the tip position of randomly excited microscopy. Second, unscented Kalman filtering based model predictive controller is proposed for the accurate tracking of the tip-position. To prevent the disadvantage of the model-based controller design, an uncertainty or unknown fault function of the system is estimated by unscented Kalman filter such that the unmodeled dynamics of the system are compensated while the control signal is produced. Note that the controller voltage being applied to the microscopy is produced based on the estimated states and parameters of the atomic force microscopy. The numerical applications present that satisfactory tracking performance for tip position is obtained by the proposed fault tolerant controller such that extended Kalman filtering-based tracking results are also compared and discussed.