Generalized part family formation through fuzzy self-organizing feature map neural network

Abstract

Group technology (GT) addresses the problem of the part family formation. Similar parts, based on a certain similarity of characteristics, are grouped into a family. A design engineer facing the task of developing a new part can use a GT code or an image of the part to determine whether similar parts exist in a computer aided design (CAD) database. The manufacturing engineer can design the cellular manufacturing system based on different families. These can dramatically shorten both the design and the manufacturing life cycle. However, owing to some unavoidable factors, like brightness of light and shift of the part, the crisp network cannot recognize the parts correctly under the above-mentioned conditions. Thus, the present study is dedicated to developing a novel fuzzy neural network (FNN) for clustering the parts into several families based on the image captured from the vision sensor. The proposed network, which possesses the fuzzy inputs as well the fuzzy weights, integrates the self-organizing feature map (SOM) neural network and the fuzzy set theory. The model evaluation results showed that the proposed FNN can provide a more accurate decision compared to the fuzzy c-means algorithm.