Takagi-Sugeno Fuzzy Logic Systems

The achievements obtained by Fuzzy Logic undoubtedly changed the way expert information is represented, manipulated, and interpreted in computational systems. Nevertheless, the initialization of Mamdani FLSs’ main parameters, namely its membership functions and their interdependency relations, is a process that depends on the knowledge of an expert (which may be subjective and is ultimately limited by its know-how). Takagi and Sugeno [1] were among the first researchers who recognized that FLSs could be further enhanced with autonomous learning techniques. Together, they proposed a new structure for the consequent part of the rules, introducing also methodologies to autonomously create and improve the FLSs’ performance. Their method uses heuristic and non-linear optimization algorithms for the antecedent part of the rule-base and a Kalman Filter for the consequent one. It is however for their innovative FLS’s structure supporting their work that Takagi and Sugeno are nowadays known in FLSs’ literature (effectively coining the concept of Takagi-Sugeno FLSs), serving their work as the stepping stone for many successful research topics.