ULTRA SONIC motor is a newly developed motor, and it has excellent performance and many useful features, e.g.: high-torque density, low speed, compactness in size, no electromagnetic interferences, and so on. USM is a kind of piezo motor. The proposed speed control scheme is assumed for these applications because they require quick and precise speed control of actuators for various load conditions. A speed control method of USM using adaptive control is proposed.
Artificial Neural Network (ANN), which follows the biological neural cells in brain, consists of a number of neurons and weighted links, and it has a good potential for control applications because it can approximate the non-linear input-output mapping of the plant. Accordingly, ANN has been applied widely in the field of power electronics control. In this paper a three layer neural network for speed controller is adopted, and the weights of the links are updated using the general back propagation in order to reduce the speed error at each sampling period. In general, the speed of USM can be controlled by driving frequency, applied voltage, and phase difference of applied voltages. This paper adopts the driving frequency as control input in order to simplify the drive system.