2021/10/29 by Shahin Ghazvineh, Gholamreza Nouri, Ghazvineh, Shahin +8
Engineering · #Computer Vision and Pattern Recognition (cs.CV) #FOS: Computer and information sciences #Infrastructure Maintenance and Monitoring #Machine Fault Diagnosis Techniques #Machine Learning (cs.LG) #Structural Health Monitoring Techniques
paper · pdf · doi:10.48550/arxiv.2110.15895
openalex publication_date 2021/10/29 · openalex created_date 2022/10/06 · openalex updated_date 2026/07/28
In this paper, we present an application of 2-D convolutional neural networks\n(2-D CNNs) designed to perform both feature extraction and classification\nstages as a single organism to solve the highlighted problems. The method uses\na network of lighted CNNs instead of deep and takes raw acceleration signals as\ninput. Using lighted CNNs, in which every one of them is optimized for a\nspecific element, increases the accuracy and makes the network faster to\nperform. Also, a new framework is proposed for decreasing the data required in\nthe training phase. We verified our method on Qatar University Grandstand\nSimulator (QUGS) benchmark data provided by Structural Dynamics Team. The\nresults showed improved accuracy over other methods, and running time was\nadequate for real-time applications.\n