Real-time position error detecting in nanomanipulation using Kalman filter

The main roadblock to Atomic Force Microscope (AFM) based nanomanipulation is lack of real time visual feedback. Although the model based visual feedback can partly solve this problem, due to the complication of nano environment, it is difficult to accurately describe the behavior of nano-objects with a model. The modeling error will lead to an inaccurate feedback and a failed manipulation. In this paper, a Kalman filter is developed to real time detect this modeling error. During manipulation, the residual between the estimated behavior and the visual display behavior is real time updated. The residual's Mahalanobis distance is calculated and compared with an threshold to determine whether there is a position error. Once the threshold is exceeded, an alarm signal will be triggered to tell the system there is a position error. Furthermore, the position error can be on-line corrected by local scan method. With the assistance of Kalman filter and local scan, the position error not only can be real-time detected, but also can be online corrected. The visual display keeps matching with the real manipulation result during the whole manipulation process, which significantly improve the efficiency of the AFM based nano-assembly. Experiments of manipulating nano-particles are presented to verify the effectiveness of Kalman filter and local scan method.