Univiariate signal preprocessing methodology for fault detection in semiconductor manufacturing process

Many studies using sensor signals have been conducted in the field of fault detection and classification (FDC) for semiconductor manufacturing processes. This is because sensor signals generated in the semiconductor production process provide important information for predicting quality and yield of the finished product. However, as the process becomes more sophisticated and refined, normal and abnormal data with similar shape appears. They only show delicate differences and it is difficult to classify them using general classification algorithms. The purpose of this research is to present a preprocessing methodology for improving classification performance. The methodology consists of four steps based on signal segmentation and clustering methods. The experimental results illustrate the better performance of the proposed procedure.