A Fusion Measurement Method Based on Kalman Filter with Improved State Block and Neural Network for Nanometer Displacement

摘要

In the field of nanomanipulation, measurement method for the nano-scale displacement is one of the key technologies, and there are many restrictions on the sensors mounted in microscopic device. In order to reduce the influence of sensors on workspace, we fused the measurements of self-sensing and time-digit-conversion (TDC) method to estimate the measured nanometer displacement. Both of the two methods have a simple measurement circuit and are easy to be integrated. They can reduce the effect of thermal radiation on workspace and work in a vacuum environment (such as SEM chamber). We proposed a method based on Kalman filter with improved state block and neural network to obtain the fusion estimation. Our method achieved a sampling rate equal to that of self-sensing, as well as a precision higher than those of the two source methods. The linearity (R2) of our method is 0.9999915 throughout 8 μ$m$range. Finally, we compared our method with the traditional fusion method based on statistics.