스마트 제조 환경에서 인공지능 기반 불량 제품 검출에 관한 연구

초록

This study investigates the application of artificial intelligence in a smart manufacturing environment to analyze sensor data collected from semiconductor processes and effectively detect defective products. The primary objective of this research is to enhance the efficiency of the manufacturing process and improve the quality of the produced goods. To achieve this, data provided by Samsung's smart factory was utilized, and various preprocessing techniques such as dimension reduction, sampling, and scaling were applied. The focus was on reducing the complexity of the data while minimizing any potential performance degradation. The results of the study confirmed that appropriate dimension reduction could decrease data complexity and improve efficiency without compromising performance. Furthermore, the research demonstrated a significant improvement in defect prediction performance, achieving an approximately 3% increase in the Geometric Mean (GM) compared to previous studies. These findings underscore the critical role that AI-driven data analysis can play in advancing smart manufacturing environments.

키워드