An in-process Pokayoke (IP) system has been developed in unmanned manufacturing cells (UMCs) to approach a zero defect rate based on fuzzy systems and neural networks approaches. The IP system is a real-time approach to detect a tooling defect in an UMC. The IP system consists of two components: (1) the fuzzy-nets classifier (FNC), which maps a state vector into a recommended action using fuzzy pattern recognition, and (2) the fuzzy-nets adaptor (FNA), which maps a state vector and it failure signal into a scalar grade that indicates state integrity. The FNA also produces the output active value, p, to upgrade FNS mapping according to the variation of the input state. The performance of the IP system was examined for an end milling operation.
展开▼
