Roller Vibration Detection with Radar Sensor and Linear Discriminant Analysis

This article is dedicated to research on the problems of quality control of rotation of industrial rollers using radar equipment. The radar signals reflected from roller are analyzed using short-time Fourier transform and classified using linear discriminant analysis technique. To evaluate the proposed approach an testbed has been implemented. To obtain a dataset a set of radar signals has been recorded. The results of experiments shown that the accuracy of the proposed approach of detecting roller condition is reach 98%.