A self-tuning AFM imaging method based on data-driven mechanism

When an atomic force microscope (AFM) is utilized to scan a sample, it is usually difficult to tune and seek out suitable control gains to achieve good imaging performance, which has been a great challenge for inexperienced operators. This paper addresses this problem, and it presents an AFM imaging method with the feat of control gains self-tuning by the utilization of some data-driven mechanism. Specifically, the CARIMA model is introduced to describe the linearized AFM system, whose parameters are identified by the data-driven technology. Then, the GPC-based optimization method is adopted to calculate PI controller parameters on-line, which yields an AFM imaging method with control gains self-tuning mechanism. Finally some verification simulation is implemented, whose result demonstrates that when the scanning speed is changed or the control parameters are chosen improperly, the proposed method adjusts the control gains automatically to reduce the control error and improve the imaging accuracy.