M.S. student Yeongtak Oh from the System Health and Risk Management Lab (advisor: Prof. Byeng Dong Youn) of SNU Dept. of Mechanical and Aerospace Engineering received Best Paper Award at the 2019 International Conference on Materials and Reliability (ICMR), for his paper titled "A Novel Fault Detection Method of Industrial Robots Using Motor Current Signals via Convolutional Neural Network (CNN).”

In the research, Mr. Oh proposed the possibility of fault diagnosis based on ripple signals extracted from the preprocessing of current signal acquired from the controlling stage of a robot. Using convolutional neural network (CNN), the research demonstrated that fault diagnosis was possible under conditions of varying temperatures and speed. It also showed how the deep learning model, which is pre-trained via grad-CAM method for analyticity, detects fault by examining certain parts of the signal. The proposed method is an excellent example of using deep learning for fault detection of robots and has potential to be applied to data acquired from actual industry environments.