Brown, M. and Harris, C.J.
Is Neurofuzzy logic a better alternative for on-line control?
Information Technology Awareness in Engineering Initiative
Full text not available from this repository.
Since the mid-eighties a lot of interest has been generated in the field of learning systems due to a resurgance of interest in such topics as: neural networks, fuzzy logic, genetic algorithms and inductive learning. Engineers are interested in these techniques in order to improve the quality of their product, to reduce the design time and to improve their flexibility. In particular, control engineers want algorithms that can adapt to new situations, being sufficiently flexible so that they can be applied to a wide range of plants and operating in a robust fashion. Conventional control algorithms have been mainly used for time-invariant, linear models, and it is hoped that the current research into learning systems can be applied to many difficult control problems. This paper will describe the theory behind some of the current adaptive neurofuzzy algorithms, drawing in comparisons with Baysian rule bases and inductive learning techniques, and outline their potential (problems) for being applied to real-world problems.
Actions (login required)