| Atomic force microscopy (AFM)-based systems are the key enabling tool in emerging nanotechnologies, such as high-density data storage devices, semiconductor lithography and nanosurgery. By using piezo-based positioners (actuators), the AFM-probe tip can be moved relative to the sample surface for observing, manipulating and fabricating objects at the nanometer scale. However, a critical problem in AFM is nano-precision control of the piezo positioner. In particular, hysteresis (as well as creep and vibration) in piezos makes precise positioning a challenge and the relatively large tracking error due to hysteresis, which is substantially larger than 100 nm, is not sufficient for emerging nanoscale applications. This thesis solves an iterative learning control (ILC) problem for hysteretic systems to achieve nano-precision positioning. Specifically, an ILC algorithm is proposed and applied to compensate for hysteresis-caused positioning error in piezo-based systems, such as AFMs, and a proof of convergence, based on the Preisach hysteresis model, is presented. Moreover, the required number of iterations to achieve a desired tracking precision is quantified, and the method is experimentally evaluated on a commercial AFM system. Results show that the proposed ILC algorithm reduces the tracking error to the noise level of the sensor measurement. |
