ANFIS data driven modeling and real-time Fuzzy Logic Controller test for a Two Tanks Hydraulic System

Abstract

This paper presents a non-linear, data driven Adaptive Network based Fuzzy Inference System (ANFIS) modeling of a Two Tanks Hydraulic System (TTHS). The paper also addresses the design of a Type 1 Fuzzy Logic Controller optimized with Genetic Algorithms (GA). The controller was designed and tested in simulation with the obtained ANFIS model and validated in real-time with the actual TTHS. Obtained model shows an accurate and adequate description of the real system, useful for many applications that require a non-linear functioning representation of the TTHS. The designed controller also demonstrates excellent performance by being able to follow diverse shaped references. This work successfully demonstrates the utility of soft-computing techniques in their application to real world industrial complex systems.