New trends in recognizing experimental drives: fuzzy logic and formal language theories

Drive systems today determine the productivity and quality of industrial processes. However, they exhibit considerable complexities related with their behavior as large uncertainties at a structure and parameter levels, multidimensionality, and strong mutual interactions. With these multiplying complexities, the usual models are becoming not accurate enough. It is necessary to complement them with other information-processing techniques that allow a better recognition of their behavior. The aim of this paper is to analyze common features, and the potential, but also the drawbacks that fuzzy logic and formal language theories show when used for recognition of patterns in experimental drives. Two prototype systems are used: an electrohydraulic drive and an induction motor drive. We underline the similarities and various aspects of both recognition methodologies, despite their use on different systems. A set of experimental learning situations with critical effects on their performance are presented and discussed.