An intelligent approach for speed stability analysis of a travelling wave ultrasonic motor based on genetic k-NN algorithm

Driving voltage, driving frequency, phase difference and operating temperature are the parameters which affect the speed stability of a travelling wave ultrasonic motor (TWUSM). The weight coefficients of these parameters should be determined for the purpose of ensuring the speed stability of a TWUSM with a maximum level under different operating conditions. In this paper, a novel approach is proposed for the speed stability analysis of the TWUSM using genetic k-nearest neighbor algorithm (k-NN) and the speed stability classes of new test observations are achieved accurately. Furthermore, the genetic k-NN algorithm is compared with the classic k-NN algorithm in terms of prediction accuracy using Euclidean, Manhattan and Minkowski distance metrics. As a result of experimental studies, it is shown that the TWUSM parameters weighted by the genetic k-NN algorithm increase the speed stability of the TWUSM significantly and the genetic k-NN algorithm outperforms the classic k-NN algorithm for all of distance metrics.