부산대학교 도서관

The application of Doppler radar to classify the differences in motions related to the differences in pants is presented in this study with an aim to efficiently design comfortable pants. The types of pants that the participant was wearing when ascending a step were classified into four types, i.e., denim pants, trekking pants, half pants, and tights. The deep learning of the Doppler radar spectrogram using the convolutional neural network (CNN) achieved an accuracy 89 % of the classification accuracy. The results indicated that the motion differences related to the comfort of the pants can be detected via Doppler radar sensing.