Abstract
This paper presents a speed control system for ultrasonic motors by PI (proportional and integral) control which has auto-tuning structure based on fuzzy reasoning realized by a neural network. Since the speed characteristics of ultrasonic motor vary with temperature and applied load torque, adjusting the control gains of PI controller is necessary in order to obtain fine speed control performance. The proposed control scheme can incorporate the operator's knowledge in speed controller using fuzzy reasoning, furthermore, it can compensate the speed characteristic variations of the motor by online learning. The usefulness and validity of the proposed control scheme is examined by experiments.