Position control of ultrasonic motors using neural network

The ultrasonic motor is a newly developed motor and it has excellent performance and many useful features, therefore, the ultrasonic motor is expected for many practical applications. However, the drive principle of the ultrasonic motor is different from that of other electromagnetic type motors, and the mathematical model of the motor has not been developed yet. Furthermore, the speed characteristics of the motor hold heavy nonlinearity and they vary with drive conditions. Therefore, the precise position control of an ultrasonic motor is generally difficult. In this paper, we propose a new position control scheme for an ultrasonic motor using a neural network. The proposed controller can approximate the nonlinear input-output mappings of the motor using the neural network, and can compensate the characteristic variations by online learning based on the general backpropagation algorithm. The trained network finally makes an inverse model of the plant. The usefulness and validity of the proposed control scheme is examined in experiments