Intelligent control of semiconductor manufacturing processes

The intelligent control mechanism relies on neural networks to model quantitatively semiconductor manufacturing processes. The learning capability of neural networks provides accurate approximations to these processes. The control strategy is based on fuzzy logic for capturing the heuristic knowledge and a feedforward neural network for capturing the process models. The hardware implementation of this scheme is under development. The limitations of classical control paradigms are reviewed. The concept of intelligent control is introduced, and ways to construct hybrid fuzzy logic/neural network controllers are discussed. Fuzzy logic control is examined. The neural network control is presented. The conceptual design of intelligent manufacturing architecture and its control mechanisms are given.<
>