조성호 김동욱 박재형 박상철 2024-09 2508-4003 https://aurora.ajou.ac.kr/handle/2018.oak/37865 https://www.kci.go.kr/kciportal/ci/sereArticleSearch/ciSereArtiView.kci?sereArticleSearchBean.artiId=ART003113349 This paper proposes a data-based prediction method to predict the cycle time for robotic arm spot welding process using artificial intelligence model. Since the prediction of the cycle time of robotic arm is crucial for its utilization in documentation of an engineering plan, as it is closely associated with overall production efficiency and safety consideration. In this paper three meth- ods for estimating cycle time are introduced; 1) Estimation method based on a standard engi- neering cycle time table, 2) Cycle time calculation method using real time simulation data, and 3) Cycle time prediction method based on artificial neural network using robotic arm process information. We provide a comparative analysis of three methods, showcasing their respective performance. Additionally, through the application of the proposed approach. we elucidate the anticipated factors that can be expected to enhance overall effectiveness. Kor 한국CDE학회 인공지능 기반 SPOT 용접 사이클 타임 예측 AI Based Cycle Time Prediction for Spot Welding Article 240 3 232 한국CDE학회 논문집 29 한국CDE학회 논문집, Vol.29 No.3, pp.232-240 10.7315/CDE.2024.232 6-DOF Robot Arm Robot trajectory planning Off-line programming Deep Learning Model Cycle time prediction model Article