Fuzzy-modelling of systems with uncertain nonlinearities and its application on pneumatic drives

Uncertain non-linearities cause distinctive difficulties concerning system modelling. Nevertheless, fuzzy models are tolerant of deviations between process and model non-linearities. As the fuzzy model is composed of several local models, a mismatch of a local model has only local effects in principle. The advantageous performance of fuzzy models compared with some conventional approaches is illustrated by modelling a pneumatic drive. The particular fuzzy model structure with multidimensional membership functions simplifies the modelling procedure. Using fuzzy cluster algorithms, the prototypes are placed automatically not regarding any limitations of coordinates. This is a distinctive simplification compared with one dimensional membership functions. Because the conclusion functions provide for additional degrees of freedom functional fuzzy systems possess better approximation capabilities compared with relational fuzzy systems