Bouc-Wen model with machine learning for SISO and MIMO nano-positioning system

Abstract

Most of the application in present world have taken ‘Nano’ as a part of their near future implementation. In Nano-Positioning System (NPS) mechanical dynamics are handled at Nano scale. Designing a precision NPS is a greater challenge. Modified linearized Bouc-Wen(BW) model gives the better hysteresis, static and dynamic behavior. In this work SISO (Sıngle-input Sıngle-Output) and MIMO (Multiple-Input Multiple-Output) models of NPS are designed. Previously, system was basically designed to handle conventional macro level dynamics and the same has been extended to handle Nano scale in this work. Kalman Filtering is included to achieve very sustainable results in NPS. The NPS is designed in MATLAB/Simulink with 3 types of electrical signals. Each design is compared using various performance parameters. Machine learning process is included to suggest the appropriate input voltage level sui to obtain desired displacement.