The authors present a fuzzified cerebellar model articulation controller (CMAC) network acting as a multivariable adaptive controller featuring self-organizing association cells and the ability for self-learning required teaching signals in real time. In particular, the original CMAC has been reformulated within a framework of a simplified fuzzy control algorithm, and the associated self-learning algorithms have been developed by incorporating the schemes of competitive learning and iterative learning control into the system. The approach described here can be thought of as either a completely unsupervised fuzzy-neural control strategy or equivalently an automatic real-time knowledge acquisition scheme. The approach has been successfully applied to a problem of multivariable blood pressure control.
展开▼