A Fuzzy-Pattern-Classifier-Based Adaptive Learning Model for Sensor Fusion

The production of printing goods is laborious. Furthermore, the print quality, especially in banknotes, must be assured. It is accepted, that print defects are generated because printing parameters, also machine parameters can change unnoticed. Therefore, a combined concept for a multi-sensory learning and classification model based on new adaptive fuzzy-pattern-classifiers for data inspection is proposed. This inspection concept, which combines optical, acoustical and other machine information, comes up with a large amount of data, which leads to multivariate methods for data analysis. Multivariate methods are useful for analysis of large and complex data sets that consist of many variables measured on large numbers of physical data.