Design, fabrication and real-time neural network control of a three degrees of freedom nano-positioner

Abstract

Summary form only given as follows. A nano-metric precision three degrees of freedom (DOF) positioner is designed and fabricated. Actuation is based on piezo-electric stack actuators. Capacitive gap sensors with less than 1.0 nanometer resolution are used for position feedback. In order to design a proper closed loop controller, the open loop characteristics of the nano-positioner (hysteresis, static stiffness, drift, frequency response, and the coupling effects) are experimentally investigated. A CMAC neural network control algorithm was applied in order to provide online learning and better tracking capability compared to a standard PID control algorithm. The closed loop controller was implemented using C language on a DSP based PC bus board. The learning CMAC control provides significantly better tracking accuracy and robustness against hysteresis and drift effects.