Two-degree-of-freedom control combining machine learning and extremum seeking for fast scanning quantum dot microscopy

Scanning Quantum Dot Microscopy is a frequency modulated atomic force microscope technique, in which a sensor molecule is attached to the microscope tip. It allows to generate high resolution images at subatomic level of the electrostatic potential of nanostructures. While Scanning Quantum Dot Microscopy provides significantly increased resolution and sensitivity compared with Kelvin probe force microscopy, the scanning speed is limited, which limits usability and the maximum size of samples. We propose a two-degree-of-freedom control approach, consisting of an extremum seeking controller and a feedforward control. The feedforward signal is based on a machine learning approach, where a Gaussian process is used to capture the already scanned part of the image and compute a prediction for the next scanning lines. The proposed control approach speeds up the scanning process by a factor of about 10 and enables to scan larger areas than before.