The paper addresses the problems of complexity in fuzzy rule based systems with multiple sensor inputs. The number of fuzzy rules in this case is an exponential function of the number of inputs. Some of the existing methods for rule base reductions are reviewed and their drawbacks summarised. As an alternative, a novel methodology for complexity management in fuzzy systems is presented which is based on formal presentation techniques such as integer tables. A Matlab example is shown illustrating the presentation of a fuzzy rule base with an integer table. Finally, some future research directions are outlined within the framework of the proposed methodology.
|Number of pages||8|
|Journal||Sensors & Transducers Journal|
|Publication status||Published - Aug 2005|