Unsupervised self-organizing adaptive fuzzy controller (USOAFC)

Supervised adaptive fuzzy controllers have good learning capability. The main problem is the ability to obtain an an on-line input/output data pattern to achieve the required characteristics. A proposed solution is the design of on-line unsupervised controller that has self-tuning and self-constructing features. A controller with these characteristics is capable of handling a wide band of instability, non-linearity, ambiguity and complexity included in systems. This research develops an unsupervised Self-Tuning Adaptive Controller (USOAFC). USOAFC is a novel idea for designing a controller. USOAFC is suitable for on-line applications. USOAFC starts with some rules. USOAFC sets out to construct the necessary rules required to drive systems to achieve the required performance. USOAFC also has the ability to tune the existing rule. USOAFC is able to deal with the ambiguity, uncertainty and instability of systems. USOAFC is simple to understand and easy to apply to many processes.