Loading...
Motion Time Study with Convolutional Neural Network
Author(s)
Ji, Jirasak
Pannakkong, Warut
Tai, Pham Duc
Jeenanunta, Chawalit
Buddhakulsomsiri, Jirachai
Date Issued
January 1, 2020
Type
Conference Paper
Abstract
Manufacturing and service industries use motion time study to determine work element time and standard time for production planning and process improvement. Traditional time study is performed by human analysts with stopwatches, and therefore, is subject to uncertainties from human errors. This study proposes an automated time study model featuring a convolutional neural network. The trained model can analyze a video footage of an operation to time the work elements. The timing data are compared with reference values using a statistical analysis. The result shows effectiveness of the model in accurately and consistently estimating the work element times and standard time of the operation.
Citation
Lecture Notes in Computer Science Including Subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics, 12482 LNAI, 249-258, 2020