抄録
The input-and-output characteristic of ultrasonic motor (USM) has non-linear elements and change with a temperature rise or fluctuation of load-mass. Therefore, it is difficult to accomplish satisfactory control performance by using the conventional PID controller. In this paper, we proposed PID controller combined with neural network (NN-PID controller). In this design method, NN-PID controller's gains consist of con-stant gains of PID controller and variable gains of NN controller. Weights of NN are adjusted by back propagation method so that control error may be minimized. This method need not long learning-hour of NN. The effectiveness of the proposed design method is confirmed by experiments using the existent USM.
