Feature referenced tip localization enhanced by probability motion model for AFM based nanomanipulations

Abstract

The spatial uncertainties of tip positioning due to the nonlinearity of the PZT scanner and thermal drift hinder the further application of the AFM based nanomanipulation. This paper brings forward feature referenced tip localization enhanced by probability motion model to reduce the spatial uncertainties. An improved motion model is probabilistically built by incorporating the PI model, the creep model and the thermal drift model. For calibrating the accurate model parameters, the statistical experiments are designed and performed. Then the tip position is optimally estimated by combining with a local scan based feature sensing method. The simulation and corresponding experiments are performed to illustrate the validity and feasibility of the calibrated parameters and the algorithm. © 2011 IEEE.